Brain Tumor Detection.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cieHVVEZgelS"
      },
      "source": [
        "# Brain Tumor Detection through CNN\n",
        "\n",
        "## DA LAB Project\n",
        "## Group Members:\n",
        "\n",
        "*   M. Salman Khan\n",
        "*   M. Maaz Khattak\n",
        "*   Syed Muhammad Shahmeer\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F1SUOEtsh_lJ"
      },
      "source": [
        "## Getting Dataset\n",
        "\n",
        "Link: https://www.kaggle.com/navoneel/brain-mri-images-for-brain-tumor-detection"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "7_ajxv5kgU9p"
      },
      "outputs": [],
      "source": [
        "! mkdir ~/.kaggle\n",
        "! cp kaggle.json ~/.kaggle/\n",
        "! chmod 600 ~/.kaggle/kaggle.json"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "c1MKKHahiHzC",
        "outputId": "ace017ca-b959-4cf2-9a3e-9772087a56c7"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading brain-mri-images-for-brain-tumor-detection.zip to /content\n",
            " 60% 9.00M/15.1M [00:00<00:00, 60.7MB/s]\n",
            "100% 15.1M/15.1M [00:00<00:00, 72.2MB/s]\n"
          ]
        }
      ],
      "source": [
        "! kaggle datasets download navoneel/brain-mri-images-for-brain-tumor-detection"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "XJ9kPTTkiZD5",
        "outputId": "fac98daf-b08b-456d-d187-28841b5bcc87"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Archive:  /content/brain-mri-images-for-brain-tumor-detection.zip\n",
            "  inflating: brain_tumor_dataset/no/1 no.jpeg  \n",
            "  inflating: brain_tumor_dataset/no/10 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/11 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/12 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/13 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/14 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/15 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/17 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/18 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/19 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/2 no.jpeg  \n",
            "  inflating: brain_tumor_dataset/no/20 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/21 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/22 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/23 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/24 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/25 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/26 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/27 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/28 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/29 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/3 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/30 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/31 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/32 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/33 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/34 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/35 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/36 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/37 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/38 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/39 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/4 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/40 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/41 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/42 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/43 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/44no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/45 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/46 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/47 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/48 no.jpeg  \n",
            "  inflating: brain_tumor_dataset/no/49 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/5 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/50 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/6 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/7 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/8 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/9 no.jpg  \n",
            "  inflating: brain_tumor_dataset/no/N1.JPG  \n",
            "  inflating: brain_tumor_dataset/no/N11.jpg  \n",
            "  inflating: brain_tumor_dataset/no/N15.jpg  \n",
            "  inflating: brain_tumor_dataset/no/N16.jpg  \n",
            "  inflating: brain_tumor_dataset/no/N17.jpg  \n",
            "  inflating: brain_tumor_dataset/no/N19.JPG  \n",
            "  inflating: brain_tumor_dataset/no/N2.JPG  \n",
            "  inflating: brain_tumor_dataset/no/N20.JPG  \n",
            "  inflating: brain_tumor_dataset/no/N21.jpg  \n",
            "  inflating: brain_tumor_dataset/no/N22.JPG  \n",
            "  inflating: brain_tumor_dataset/no/N26.JPG  \n",
            "  inflating: brain_tumor_dataset/no/N3.jpg  \n",
            "  inflating: brain_tumor_dataset/no/N5.jpg  \n",
            "  inflating: brain_tumor_dataset/no/N6.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No11.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No12.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No13.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No14.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No15.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No16.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No17.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No18.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No19.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No20.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No21.jpg  \n",
            "  inflating: brain_tumor_dataset/no/No22.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 1.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 10.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 100.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 2.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 3.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 4.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 5.jpeg  \n",
            "  inflating: brain_tumor_dataset/no/no 6.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 7.jpeg  \n",
            "  inflating: brain_tumor_dataset/no/no 8.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 89.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 9.png  \n",
            "  inflating: brain_tumor_dataset/no/no 90.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 91.jpeg  \n",
            "  inflating: brain_tumor_dataset/no/no 92.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 923.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 94.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 95.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 96.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 97.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 98.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no 99.jpg  \n",
            "  inflating: brain_tumor_dataset/no/no.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y1.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y10.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y100.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y101.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y102.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y103.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y104.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y105.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y106.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y107.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y108.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y109.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y11.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y111.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y112.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y113.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y114.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y115.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y116.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y117.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y12.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y120.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y13.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y14.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y146.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y147.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y148.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y15.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y153.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y154.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y155.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y156.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y157.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y158.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y159.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y16.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y160.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y161.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y162.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y163.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y164.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y165.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y166.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y167.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y168.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y169.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y17.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y170.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y18.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y180.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y181.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y182.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y183.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y184.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y185.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y186.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y187.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y188.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y19.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y192.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y193.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y194.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y195.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y2.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y20.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y21.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y22.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y23.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y24.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y242.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y243.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y244.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y245.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y246.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y247.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y248.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y249.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y25.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y250.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y251.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y252.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y253.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y254.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y255.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y256.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y257.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y258.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y259.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y26.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y27.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y28.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y29.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y3.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y30.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y31.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y32.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y33.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y34.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y35.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y36.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y37.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y38.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y39.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y4.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y40.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y41.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y42.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y44.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y45.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y46.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y47.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y49.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y50.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y51.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y52.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y53.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y54.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y55.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y56.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y58.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y59.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y6.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y60.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y61.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y62.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y65.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y66.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y67.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y69.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y7.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y70.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y71.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y73.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y74.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y75.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y76.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y77.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y78.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y79.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y8.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y81.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y82.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y85.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y86.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y89.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y9.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y90.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y91.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y92.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y92.png  \n",
            "  inflating: brain_tumor_dataset/yes/Y95.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y96.jpg  \n",
            "  inflating: brain_tumor_dataset/yes/Y97.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y98.JPG  \n",
            "  inflating: brain_tumor_dataset/yes/Y99.JPG  \n",
            "  inflating: no/1 no.jpeg            \n",
            "  inflating: no/10 no.jpg            \n",
            "  inflating: no/11 no.jpg            \n",
            "  inflating: no/12 no.jpg            \n",
            "  inflating: no/13 no.jpg            \n",
            "  inflating: no/14 no.jpg            \n",
            "  inflating: no/15 no.jpg            \n",
            "  inflating: no/17 no.jpg            \n",
            "  inflating: no/18 no.jpg            \n",
            "  inflating: no/19 no.jpg            \n",
            "  inflating: no/2 no.jpeg            \n",
            "  inflating: no/20 no.jpg            \n",
            "  inflating: no/21 no.jpg            \n",
            "  inflating: no/22 no.jpg            \n",
            "  inflating: no/23 no.jpg            \n",
            "  inflating: no/24 no.jpg            \n",
            "  inflating: no/25 no.jpg            \n",
            "  inflating: no/26 no.jpg            \n",
            "  inflating: no/27 no.jpg            \n",
            "  inflating: no/28 no.jpg            \n",
            "  inflating: no/29 no.jpg            \n",
            "  inflating: no/3 no.jpg             \n",
            "  inflating: no/30 no.jpg            \n",
            "  inflating: no/31 no.jpg            \n",
            "  inflating: no/32 no.jpg            \n",
            "  inflating: no/33 no.jpg            \n",
            "  inflating: no/34 no.jpg            \n",
            "  inflating: no/35 no.jpg            \n",
            "  inflating: no/36 no.jpg            \n",
            "  inflating: no/37 no.jpg            \n",
            "  inflating: no/38 no.jpg            \n",
            "  inflating: no/39 no.jpg            \n",
            "  inflating: no/4 no.jpg             \n",
            "  inflating: no/40 no.jpg            \n",
            "  inflating: no/41 no.jpg            \n",
            "  inflating: no/42 no.jpg            \n",
            "  inflating: no/43 no.jpg            \n",
            "  inflating: no/44no.jpg             \n",
            "  inflating: no/45 no.jpg            \n",
            "  inflating: no/46 no.jpg            \n",
            "  inflating: no/47 no.jpg            \n",
            "  inflating: no/48 no.jpeg           \n",
            "  inflating: no/49 no.jpg            \n",
            "  inflating: no/5 no.jpg             \n",
            "  inflating: no/50 no.jpg            \n",
            "  inflating: no/6 no.jpg             \n",
            "  inflating: no/7 no.jpg             \n",
            "  inflating: no/8 no.jpg             \n",
            "  inflating: no/9 no.jpg             \n",
            "  inflating: no/N1.JPG               \n",
            "  inflating: no/N11.jpg              \n",
            "  inflating: no/N15.jpg              \n",
            "  inflating: no/N16.jpg              \n",
            "  inflating: no/N17.jpg              \n",
            "  inflating: no/N19.JPG              \n",
            "  inflating: no/N2.JPG               \n",
            "  inflating: no/N20.JPG              \n",
            "  inflating: no/N21.jpg              \n",
            "  inflating: no/N22.JPG              \n",
            "  inflating: no/N26.JPG              \n",
            "  inflating: no/N3.jpg               \n",
            "  inflating: no/N5.jpg               \n",
            "  inflating: no/N6.jpg               \n",
            "  inflating: no/No11.jpg             \n",
            "  inflating: no/No12.jpg             \n",
            "  inflating: no/No13.jpg             \n",
            "  inflating: no/No14.jpg             \n",
            "  inflating: no/No15.jpg             \n",
            "  inflating: no/No16.jpg             \n",
            "  inflating: no/No17.jpg             \n",
            "  inflating: no/No18.jpg             \n",
            "  inflating: no/No19.jpg             \n",
            "  inflating: no/No20.jpg             \n",
            "  inflating: no/No21.jpg             \n",
            "  inflating: no/No22.jpg             \n",
            "  inflating: no/no 1.jpg             \n",
            "  inflating: no/no 10.jpg            \n",
            "  inflating: no/no 100.jpg           \n",
            "  inflating: no/no 2.jpg             \n",
            "  inflating: no/no 3.jpg             \n",
            "  inflating: no/no 4.jpg             \n",
            "  inflating: no/no 5.jpeg            \n",
            "  inflating: no/no 6.jpg             \n",
            "  inflating: no/no 7.jpeg            \n",
            "  inflating: no/no 8.jpg             \n",
            "  inflating: no/no 89.jpg            \n",
            "  inflating: no/no 9.png             \n",
            "  inflating: no/no 90.jpg            \n",
            "  inflating: no/no 91.jpeg           \n",
            "  inflating: no/no 92.jpg            \n",
            "  inflating: no/no 923.jpg           \n",
            "  inflating: no/no 94.jpg            \n",
            "  inflating: no/no 95.jpg            \n",
            "  inflating: no/no 96.jpg            \n",
            "  inflating: no/no 97.jpg            \n",
            "  inflating: no/no 98.jpg            \n",
            "  inflating: no/no 99.jpg            \n",
            "  inflating: no/no.jpg               \n",
            "  inflating: yes/Y1.jpg              \n",
            "  inflating: yes/Y10.jpg             \n",
            "  inflating: yes/Y100.JPG            \n",
            "  inflating: yes/Y101.jpg            \n",
            "  inflating: yes/Y102.jpg            \n",
            "  inflating: yes/Y103.jpg            \n",
            "  inflating: yes/Y104.jpg            \n",
            "  inflating: yes/Y105.jpg            \n",
            "  inflating: yes/Y106.jpg            \n",
            "  inflating: yes/Y107.jpg            \n",
            "  inflating: yes/Y108.jpg            \n",
            "  inflating: yes/Y109.JPG            \n",
            "  inflating: yes/Y11.jpg             \n",
            "  inflating: yes/Y111.JPG            \n",
            "  inflating: yes/Y112.JPG            \n",
            "  inflating: yes/Y113.JPG            \n",
            "  inflating: yes/Y114.JPG            \n",
            "  inflating: yes/Y115.JPG            \n",
            "  inflating: yes/Y116.JPG            \n",
            "  inflating: yes/Y117.JPG            \n",
            "  inflating: yes/Y12.jpg             \n",
            "  inflating: yes/Y120.JPG            \n",
            "  inflating: yes/Y13.jpg             \n",
            "  inflating: yes/Y14.jpg             \n",
            "  inflating: yes/Y146.JPG            \n",
            "  inflating: yes/Y147.JPG            \n",
            "  inflating: yes/Y148.JPG            \n",
            "  inflating: yes/Y15.jpg             \n",
            "  inflating: yes/Y153.jpg            \n",
            "  inflating: yes/Y154.jpg            \n",
            "  inflating: yes/Y155.JPG            \n",
            "  inflating: yes/Y156.JPG            \n",
            "  inflating: yes/Y157.JPG            \n",
            "  inflating: yes/Y158.JPG            \n",
            "  inflating: yes/Y159.JPG            \n",
            "  inflating: yes/Y16.JPG             \n",
            "  inflating: yes/Y160.JPG            \n",
            "  inflating: yes/Y161.JPG            \n",
            "  inflating: yes/Y162.jpg            \n",
            "  inflating: yes/Y163.JPG            \n",
            "  inflating: yes/Y164.JPG            \n",
            "  inflating: yes/Y165.JPG            \n",
            "  inflating: yes/Y166.JPG            \n",
            "  inflating: yes/Y167.JPG            \n",
            "  inflating: yes/Y168.jpg            \n",
            "  inflating: yes/Y169.jpg            \n",
            "  inflating: yes/Y17.jpg             \n",
            "  inflating: yes/Y170.JPG            \n",
            "  inflating: yes/Y18.JPG             \n",
            "  inflating: yes/Y180.jpg            \n",
            "  inflating: yes/Y181.jpg            \n",
            "  inflating: yes/Y182.JPG            \n",
            "  inflating: yes/Y183.jpg            \n",
            "  inflating: yes/Y184.JPG            \n",
            "  inflating: yes/Y185.jpg            \n",
            "  inflating: yes/Y186.jpg            \n",
            "  inflating: yes/Y187.jpg            \n",
            "  inflating: yes/Y188.jpg            \n",
            "  inflating: yes/Y19.JPG             \n",
            "  inflating: yes/Y192.JPG            \n",
            "  inflating: yes/Y193.JPG            \n",
            "  inflating: yes/Y194.jpg            \n",
            "  inflating: yes/Y195.JPG            \n",
            "  inflating: yes/Y2.jpg              \n",
            "  inflating: yes/Y20.jpg             \n",
            "  inflating: yes/Y21.jpg             \n",
            "  inflating: yes/Y22.jpg             \n",
            "  inflating: yes/Y23.JPG             \n",
            "  inflating: yes/Y24.jpg             \n",
            "  inflating: yes/Y242.JPG            \n",
            "  inflating: yes/Y243.JPG            \n",
            "  inflating: yes/Y244.JPG            \n",
            "  inflating: yes/Y245.jpg            \n",
            "  inflating: yes/Y246.JPG            \n",
            "  inflating: yes/Y247.JPG            \n",
            "  inflating: yes/Y248.JPG            \n",
            "  inflating: yes/Y249.JPG            \n",
            "  inflating: yes/Y25.jpg             \n",
            "  inflating: yes/Y250.jpg            \n",
            "  inflating: yes/Y251.JPG            \n",
            "  inflating: yes/Y252.jpg            \n",
            "  inflating: yes/Y253.JPG            \n",
            "  inflating: yes/Y254.jpg            \n",
            "  inflating: yes/Y255.JPG            \n",
            "  inflating: yes/Y256.JPG            \n",
            "  inflating: yes/Y257.jpg            \n",
            "  inflating: yes/Y258.JPG            \n",
            "  inflating: yes/Y259.JPG            \n",
            "  inflating: yes/Y26.jpg             \n",
            "  inflating: yes/Y27.jpg             \n",
            "  inflating: yes/Y28.jpg             \n",
            "  inflating: yes/Y29.jpg             \n",
            "  inflating: yes/Y3.jpg              \n",
            "  inflating: yes/Y30.jpg             \n",
            "  inflating: yes/Y31.jpg             \n",
            "  inflating: yes/Y32.jpg             \n",
            "  inflating: yes/Y33.jpg             \n",
            "  inflating: yes/Y34.jpg             \n",
            "  inflating: yes/Y35.jpg             \n",
            "  inflating: yes/Y36.JPG             \n",
            "  inflating: yes/Y37.jpg             \n",
            "  inflating: yes/Y38.jpg             \n",
            "  inflating: yes/Y39.jpg             \n",
            "  inflating: yes/Y4.jpg              \n",
            "  inflating: yes/Y40.JPG             \n",
            "  inflating: yes/Y41.jpg             \n",
            "  inflating: yes/Y42.jpg             \n",
            "  inflating: yes/Y44.JPG             \n",
            "  inflating: yes/Y45.JPG             \n",
            "  inflating: yes/Y46.jpg             \n",
            "  inflating: yes/Y47.JPG             \n",
            "  inflating: yes/Y49.JPG             \n",
            "  inflating: yes/Y50.JPG             \n",
            "  inflating: yes/Y51.jpg             \n",
            "  inflating: yes/Y52.jpg             \n",
            "  inflating: yes/Y53.jpg             \n",
            "  inflating: yes/Y54.jpg             \n",
            "  inflating: yes/Y55.jpg             \n",
            "  inflating: yes/Y56.jpg             \n",
            "  inflating: yes/Y58.JPG             \n",
            "  inflating: yes/Y59.JPG             \n",
            "  inflating: yes/Y6.jpg              \n",
            "  inflating: yes/Y60.jpg             \n",
            "  inflating: yes/Y61.jpg             \n",
            "  inflating: yes/Y62.jpg             \n",
            "  inflating: yes/Y65.JPG             \n",
            "  inflating: yes/Y66.JPG             \n",
            "  inflating: yes/Y67.JPG             \n",
            "  inflating: yes/Y69.jpg             \n",
            "  inflating: yes/Y7.jpg              \n",
            "  inflating: yes/Y70.jpg             \n",
            "  inflating: yes/Y71.JPG             \n",
            "  inflating: yes/Y73.jpg             \n",
            "  inflating: yes/Y74.jpg             \n",
            "  inflating: yes/Y75.JPG             \n",
            "  inflating: yes/Y76.jpg             \n",
            "  inflating: yes/Y77.jpg             \n",
            "  inflating: yes/Y78.jpg             \n",
            "  inflating: yes/Y79.jpg             \n",
            "  inflating: yes/Y8.jpg              \n",
            "  inflating: yes/Y81.jpg             \n",
            "  inflating: yes/Y82.jpg             \n",
            "  inflating: yes/Y85.JPG             \n",
            "  inflating: yes/Y86.JPG             \n",
            "  inflating: yes/Y89.JPG             \n",
            "  inflating: yes/Y9.jpg              \n",
            "  inflating: yes/Y90.jpg             \n",
            "  inflating: yes/Y91.jpg             \n",
            "  inflating: yes/Y92.jpg             \n",
            "  inflating: yes/Y92.png             \n",
            "  inflating: yes/Y95.jpg             \n",
            "  inflating: yes/Y96.jpg             \n",
            "  inflating: yes/Y97.JPG             \n",
            "  inflating: yes/Y98.JPG             \n",
            "  inflating: yes/Y99.JPG             \n"
          ]
        }
      ],
      "source": [
        "! unzip /content/brain-mri-images-for-brain-tumor-detection.zip"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QEo6RSgDieW7"
      },
      "source": [
        "## Importing Libraries and Dependencies"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "7XaUogxWiswl"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator,array_to_img\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import seaborn as sns\n",
        "import sklearn\n",
        "from sklearn.metrics import confusion_matrix"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gK2cZDHfjdBZ"
      },
      "source": [
        "### EDA"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "4boMeEXMi_s8",
        "outputId": "9b70f9f2-0d54-4b5f-942a-1476fe4e2a81"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "2"
            ]
          },
          "execution_count": 8,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "len(os.listdir(\"/content/brain_tumor_dataset/\"))\n",
        "# 2 Classes"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "qFqQ0Uo7jrm7"
      },
      "outputs": [],
      "source": [
        "train_dir = \"/content/brain_tumor_dataset\""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YhmmHpspkVGQ"
      },
      "source": [
        "## Performing Image Augmentation for better generalization plus test-valid split"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "Gji-I_vYkCJt",
        "outputId": "e8ea388a-87b3-4d4e-a9d4-5706400d5cea"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Found 203 images belonging to 2 classes.\n",
            "Found 50 images belonging to 2 classes.\n"
          ]
        }
      ],
      "source": [
        "train_datagen = ImageDataGenerator(rescale=1./255,\n",
        "                                     rotation_range=10,\n",
        "                                     width_shift_range=0.05,\n",
        "                                     height_shift_range=0.05,\n",
        "                                     zoom_range=0.05,\n",
        "                                     horizontal_flip=True, \n",
        "                                     vertical_flip=True,\n",
        "                                     brightness_range = [0.5, 1.4],\n",
        "                                     validation_split=0.2)\n",
        "\n",
        "train_set = train_datagen.flow_from_directory(directory=train_dir,\n",
        "                                                    batch_size=10,\n",
        "                                                    target_size=(224, 224),\n",
        "                                                    shuffle = True,\n",
        "                                                    class_mode = \"binary\",\n",
        "                                                    subset = \"training\")\n",
        "\n",
        "valid_set = train_datagen.flow_from_directory(directory=train_dir,\n",
        "                                                            batch_size=10,\n",
        "                                                            target_size=(224, 224),\n",
        "                                                            shuffle = False,\n",
        "                                                            class_mode = \"binary\",\n",
        "                                                            subset = \"validation\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KTRNtFvZk321"
      },
      "source": [
        "## Plotting Augmented Images"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {
        "id": "pJ-bb06llGNo"
      },
      "outputs": [],
      "source": [
        "x, label = train_set.next()\n",
        "def plot_images(x, label):\n",
        "  for im, l in zip(x, label):\n",
        "    plt.figure(figsize=(3, 3))\n",
        "    plt.xlabel(l)\n",
        "    plt.imshow(im)\n",
        "    plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 653
        },
        "id": "6wB8zvcJoPYt",
        "outputId": "8d3687dd-df31-4474-87d7-39d9199e761a"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 216x216 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 216x216 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 216x216 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_images(x[:3], label[:3])\n",
        "# 1.0 label for 'yes' and 0.0 for 'no'"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H6PtfbKdqHEq"
      },
      "source": [
        "## Creating Simple CNN Model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "id": "CG2e_yVpqtyV"
      },
      "outputs": [],
      "source": [
        "model = tf.keras.models.Sequential([ \n",
        "      tf.keras.layers.Conv2D(filters = 64 , kernel_size = (3,3), activation=\"relu\", input_shape=(224,224,3)),\n",
        "      tf.keras.layers.MaxPooling2D(pool_size = (3,3)),\n",
        "      tf.keras.layers.Conv2D(filters = 32 , kernel_size = (3,3), activation=\"relu\"),\n",
        "      tf.keras.layers.MaxPooling2D(pool_size = (3,3)),\n",
        "      tf.keras.layers.Conv2D(filters = 32 , kernel_size = (3,3), activation=\"relu\"),\n",
        "      tf.keras.layers.MaxPooling2D(pool_size = (3,3)),\n",
        "      tf.keras.layers.Flatten(),\n",
        "      tf.keras.layers.Dense(units = 512 , activation=\"relu\"),\n",
        "      tf.keras.layers.Dropout(rate = 0.5),\n",
        "      tf.keras.layers.Dense(units = 1 , activation=\"sigmoid\")\n",
        "  ])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "tZDifJi5rNrM",
        "outputId": "44dab311-b13d-48bf-f134-c27bc089e166"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d (Conv2D)             (None, 222, 222, 64)      1792      \n",
            "                                                                 \n",
            " max_pooling2d (MaxPooling2D  (None, 74, 74, 64)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 72, 72, 32)        18464     \n",
            "                                                                 \n",
            " max_pooling2d_1 (MaxPooling  (None, 24, 24, 32)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_2 (Conv2D)           (None, 22, 22, 32)        9248      \n",
            "                                                                 \n",
            " max_pooling2d_2 (MaxPooling  (None, 7, 7, 32)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 1568)              0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 512)               803328    \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 512)               0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 1)                 513       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 833,345\n",
            "Trainable params: 833,345\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "metrics = [\n",
        "    tf.keras.metrics.TruePositives(name='tp'),tf.keras.metrics.FalsePositives(name='fp'), # predict postive(1) for negative(0)\n",
        "    tf.keras.metrics.TrueNegatives(name='tn'),tf.keras.metrics.FalseNegatives(name='fn'), # predict negative(0) for postive(1)\n",
        "    tf.keras.metrics.BinaryAccuracy(name='accuracy'),tf.keras.metrics.Precision(name='precision'),\n",
        "]\n",
        "\n",
        "model.compile(optimizer = tf.optimizers.Adam(learning_rate = 0.0001),\n",
        "                loss = \"binary_crossentropy\",\n",
        "                metrics=metrics)\n",
        "\n",
        "model.summary()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uM3ZVEL1tzRG"
      },
      "source": [
        "## Training"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "GD92YRpvsCAB",
        "outputId": "43a0de0d-1043-445f-8aa5-2f9bf45bf458"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "21/21 [==============================] - 18s 729ms/step - loss: 0.6606 - tp: 122.0000 - fp: 77.0000 - tn: 2.0000 - fn: 2.0000 - accuracy: 0.6108 - precision: 0.6131 - val_loss: 0.6262 - val_tp: 31.0000 - val_fp: 18.0000 - val_tn: 1.0000 - val_fn: 0.0000e+00 - val_accuracy: 0.6400 - val_precision: 0.6327\n",
            "Epoch 2/20\n",
            "21/21 [==============================] - 15s 689ms/step - loss: 0.6291 - tp: 123.0000 - fp: 56.0000 - tn: 23.0000 - fn: 1.0000 - accuracy: 0.7192 - precision: 0.6872 - val_loss: 0.5724 - val_tp: 30.0000 - val_fp: 15.0000 - val_tn: 4.0000 - val_fn: 1.0000 - val_accuracy: 0.6800 - val_precision: 0.6667\n",
            "Epoch 3/20\n",
            "21/21 [==============================] - 15s 688ms/step - loss: 0.5852 - tp: 115.0000 - fp: 47.0000 - tn: 32.0000 - fn: 9.0000 - accuracy: 0.7241 - precision: 0.7099 - val_loss: 0.5176 - val_tp: 29.0000 - val_fp: 10.0000 - val_tn: 9.0000 - val_fn: 2.0000 - val_accuracy: 0.7600 - val_precision: 0.7436\n",
            "Epoch 4/20\n",
            "21/21 [==============================] - 15s 690ms/step - loss: 0.5594 - tp: 105.0000 - fp: 39.0000 - tn: 40.0000 - fn: 19.0000 - accuracy: 0.7143 - precision: 0.7292 - val_loss: 0.5575 - val_tp: 20.0000 - val_fp: 7.0000 - val_tn: 12.0000 - val_fn: 11.0000 - val_accuracy: 0.6400 - val_precision: 0.7407\n",
            "Epoch 5/20\n",
            "21/21 [==============================] - 15s 696ms/step - loss: 0.5536 - tp: 109.0000 - fp: 36.0000 - tn: 43.0000 - fn: 15.0000 - accuracy: 0.7488 - precision: 0.7517 - val_loss: 0.5649 - val_tp: 27.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 4.0000 - val_accuracy: 0.7400 - val_precision: 0.7500\n",
            "Epoch 6/20\n",
            "21/21 [==============================] - 15s 696ms/step - loss: 0.5164 - tp: 104.0000 - fp: 29.0000 - tn: 50.0000 - fn: 20.0000 - accuracy: 0.7586 - precision: 0.7820 - val_loss: 0.5090 - val_tp: 26.0000 - val_fp: 7.0000 - val_tn: 12.0000 - val_fn: 5.0000 - val_accuracy: 0.7600 - val_precision: 0.7879\n",
            "Epoch 7/20\n",
            "21/21 [==============================] - 15s 696ms/step - loss: 0.4980 - tp: 113.0000 - fp: 31.0000 - tn: 48.0000 - fn: 11.0000 - accuracy: 0.7931 - precision: 0.7847 - val_loss: 0.5200 - val_tp: 28.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 3.0000 - val_accuracy: 0.7600 - val_precision: 0.7568\n",
            "Epoch 8/20\n",
            "21/21 [==============================] - 15s 694ms/step - loss: 0.4948 - tp: 108.0000 - fp: 33.0000 - tn: 46.0000 - fn: 16.0000 - accuracy: 0.7586 - precision: 0.7660 - val_loss: 0.6220 - val_tp: 23.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 8.0000 - val_accuracy: 0.6600 - val_precision: 0.7188\n",
            "Epoch 9/20\n",
            "21/21 [==============================] - 15s 695ms/step - loss: 0.4822 - tp: 108.0000 - fp: 26.0000 - tn: 53.0000 - fn: 16.0000 - accuracy: 0.7931 - precision: 0.8060 - val_loss: 0.5681 - val_tp: 27.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 4.0000 - val_accuracy: 0.7400 - val_precision: 0.7500\n",
            "Epoch 10/20\n",
            "21/21 [==============================] - 15s 698ms/step - loss: 0.4977 - tp: 100.0000 - fp: 27.0000 - tn: 52.0000 - fn: 24.0000 - accuracy: 0.7488 - precision: 0.7874 - val_loss: 0.5569 - val_tp: 28.0000 - val_fp: 10.0000 - val_tn: 9.0000 - val_fn: 3.0000 - val_accuracy: 0.7400 - val_precision: 0.7368\n",
            "Epoch 11/20\n",
            "21/21 [==============================] - 15s 696ms/step - loss: 0.5065 - tp: 111.0000 - fp: 25.0000 - tn: 54.0000 - fn: 13.0000 - accuracy: 0.8128 - precision: 0.8162 - val_loss: 0.5256 - val_tp: 27.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 4.0000 - val_accuracy: 0.7400 - val_precision: 0.7500\n",
            "Epoch 12/20\n",
            "21/21 [==============================] - 15s 720ms/step - loss: 0.5125 - tp: 111.0000 - fp: 27.0000 - tn: 52.0000 - fn: 13.0000 - accuracy: 0.8030 - precision: 0.8043 - val_loss: 0.4954 - val_tp: 28.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 3.0000 - val_accuracy: 0.7600 - val_precision: 0.7568\n",
            "Epoch 13/20\n",
            "21/21 [==============================] - 15s 690ms/step - loss: 0.4931 - tp: 108.0000 - fp: 24.0000 - tn: 55.0000 - fn: 16.0000 - accuracy: 0.8030 - precision: 0.8182 - val_loss: 0.5591 - val_tp: 28.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 3.0000 - val_accuracy: 0.7600 - val_precision: 0.7568\n",
            "Epoch 14/20\n",
            "21/21 [==============================] - 15s 692ms/step - loss: 0.4668 - tp: 114.0000 - fp: 32.0000 - tn: 47.0000 - fn: 10.0000 - accuracy: 0.7931 - precision: 0.7808 - val_loss: 0.5332 - val_tp: 27.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 4.0000 - val_accuracy: 0.7400 - val_precision: 0.7500\n",
            "Epoch 15/20\n",
            "21/21 [==============================] - 15s 699ms/step - loss: 0.4973 - tp: 105.0000 - fp: 27.0000 - tn: 52.0000 - fn: 19.0000 - accuracy: 0.7734 - precision: 0.7955 - val_loss: 0.5854 - val_tp: 29.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 2.0000 - val_accuracy: 0.7800 - val_precision: 0.7632\n",
            "Epoch 16/20\n",
            "21/21 [==============================] - 15s 687ms/step - loss: 0.4878 - tp: 110.0000 - fp: 27.0000 - tn: 52.0000 - fn: 14.0000 - accuracy: 0.7980 - precision: 0.8029 - val_loss: 0.5427 - val_tp: 28.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 3.0000 - val_accuracy: 0.7600 - val_precision: 0.7568\n",
            "Epoch 17/20\n",
            "21/21 [==============================] - 15s 691ms/step - loss: 0.4647 - tp: 110.0000 - fp: 24.0000 - tn: 55.0000 - fn: 14.0000 - accuracy: 0.8128 - precision: 0.8209 - val_loss: 0.5729 - val_tp: 26.0000 - val_fp: 8.0000 - val_tn: 11.0000 - val_fn: 5.0000 - val_accuracy: 0.7400 - val_precision: 0.7647\n",
            "Epoch 18/20\n",
            "21/21 [==============================] - 15s 697ms/step - loss: 0.4543 - tp: 109.0000 - fp: 23.0000 - tn: 56.0000 - fn: 15.0000 - accuracy: 0.8128 - precision: 0.8258 - val_loss: 0.5554 - val_tp: 28.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 3.0000 - val_accuracy: 0.7600 - val_precision: 0.7568\n",
            "Epoch 19/20\n",
            "21/21 [==============================] - 15s 701ms/step - loss: 0.4736 - tp: 106.0000 - fp: 24.0000 - tn: 55.0000 - fn: 18.0000 - accuracy: 0.7931 - precision: 0.8154 - val_loss: 0.5586 - val_tp: 28.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 3.0000 - val_accuracy: 0.7600 - val_precision: 0.7568\n",
            "Epoch 20/20\n",
            "21/21 [==============================] - 15s 691ms/step - loss: 0.4838 - tp: 114.0000 - fp: 25.0000 - tn: 54.0000 - fn: 10.0000 - accuracy: 0.8276 - precision: 0.8201 - val_loss: 0.4966 - val_tp: 26.0000 - val_fp: 9.0000 - val_tn: 10.0000 - val_fn: 5.0000 - val_accuracy: 0.7200 - val_precision: 0.7429\n"
          ]
        }
      ],
      "source": [
        "history1 = model.fit(train_set, epochs = 20, validation_data = valid_set)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "uQVZ_pi-tcSn",
        "outputId": "5288f4f6-2334-4f58-b18d-2a54ee69dfb1"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "loss :  0.5671152472496033\n",
            "tp :  27.0\n",
            "fp :  9.0\n",
            "tn :  10.0\n",
            "fn :  4.0\n",
            "accuracy :  0.7400000095367432\n",
            "precision :  0.75\n"
          ]
        }
      ],
      "source": [
        "validation_results = model.evaluate(valid_set, verbose=0)\n",
        "\n",
        "for name, value in zip(model.metrics_names, validation_results):\n",
        "  print(name, ': ', value)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 38,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 437
        },
        "id": "shlD8ePvrpZV",
        "outputId": "64fd3018-f16b-4117-99ac-594d7ab11f15"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Non-tumor detected (True Negatives):  11\n",
            "Non-tumor incorrectly detected (False Positives):  8\n",
            "Tumor missed (False Negatives):  3\n",
            "Tumor detected (True Positives):  28\n",
            "Total case:  31\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 360x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "validation_predictions = model.predict(valid_set)\n",
        "\n",
        "labels = valid_set.labels\n",
        "\n",
        "cm = confusion_matrix(labels, validation_predictions > 0.5)\n",
        "plt.figure(figsize=(5,5))\n",
        "sns.heatmap(cm, annot=True, fmt=\"d\")\n",
        "plt.title('Confusion matrix @{:.2f}'.format(0.5))\n",
        "plt.ylabel('Actual label')\n",
        "plt.xlabel('Predicted label')\n",
        "\n",
        "print('Non-tumor detected (True Negatives): ', cm[0][0])\n",
        "print('Non-tumor incorrectly detected (False Positives): ', cm[0][1])\n",
        "print('Tumor missed (False Negatives): ', cm[1][0])\n",
        "print('Tumor detected (True Positives): ', cm[1][1])\n",
        "print('Total case: ', np.sum(cm[1]))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a3PTP0hAw7pS"
      },
      "source": [
        "## Results"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 45,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "r0UMZunjxypU",
        "outputId": "32780fae-ff23-461e-ec84-b23bfdc3f1e9"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "dict_keys(['loss', 'tp', 'fp', 'tn', 'fn', 'accuracy', 'precision', 'val_loss', 'val_tp', 'val_fp', 'val_tn', 'val_fn', 'val_accuracy', 'val_precision'])"
            ]
          },
          "execution_count": 45,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "h = history1.history\n",
        "h.keys()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 48,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 312
        },
        "id": "jwD_HWirw-Jw",
        "outputId": "e22d109f-454a-4c5f-d13a-88602dffce1b"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<function matplotlib.pyplot.show>"
            ]
          },
          "execution_count": 48,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.plot(h['accuracy'])\n",
        "plt.plot(h['val_accuracy'], c='red')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('%')\n",
        "plt.title('acc vs val_acc')\n",
        "plt.show"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 49,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 312
        },
        "id": "C-tvKbFMyi1e",
        "outputId": "8d434aff-a647-48d4-e5fa-cb37413f2df8"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<function matplotlib.pyplot.show>"
            ]
          },
          "execution_count": 49,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.plot(h['accuracy'])\n",
        "plt.plot(h['loss'], c='red')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('%')\n",
        "plt.title('acc vs loss')\n",
        "plt.show"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UEO16iW8zYeX"
      },
      "source": [
        "## Transfer Learning using MobileNet V2"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 50,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "rCTqmdRQzdkX",
        "outputId": "9a5a1366-1dff-4ce0-d6d9-f82461044b9f"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/mobilenet_v2/mobilenet_v2_weights_tf_dim_ordering_tf_kernels_1.0_224_no_top.h5\n",
            "9412608/9406464 [==============================] - 0s 0us/step\n",
            "9420800/9406464 [==============================] - 0s 0us/step\n",
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " mobilenetv2_1.00_224 (Funct  (None, 7, 7, 1280)       2257984   \n",
            " ional)                                                          \n",
            "                                                                 \n",
            " global_average_pooling2d (G  (None, 1280)             0         \n",
            " lobalAveragePooling2D)                                          \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 1024)              1311744   \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 1024)              0         \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 1)                 1025      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 3,570,753\n",
            "Trainable params: 1,312,769\n",
            "Non-trainable params: 2,257,984\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "base_model = tf.keras.applications.MobileNetV2(input_shape=(224,224,3),\n",
        "                                               include_top=False,\n",
        "                                               weights='imagenet')\n",
        "\n",
        "base_model.trainable = False\n",
        "\n",
        "model = tf.keras.Sequential([\n",
        "                              base_model,\n",
        "                              tf.keras.layers.GlobalAveragePooling2D(),\n",
        "                              tf.keras.layers.Dense(1024, activation = 'relu'),\n",
        "                              tf.keras.layers.Dropout(rate = 0.2),\n",
        "                              tf.keras.layers.Dense(1, activation='sigmoid')\n",
        "                            ])\n",
        "\n",
        "model.compile(optimizer = tf.optimizers.Adam(learning_rate = 0.00008),\n",
        "                loss = \"binary_crossentropy\",\n",
        "                metrics=metrics)\n",
        "\n",
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 51,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "0YG9fJMkzqv6",
        "outputId": "046dc77c-fa6d-42c1-c067-d19df8887dc3"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "21/21 [==============================] - 14s 513ms/step - loss: 0.6551 - tp: 134.0000 - fp: 62.0000 - tn: 36.0000 - fn: 21.0000 - accuracy: 0.6719 - precision: 0.6837 - val_loss: 0.4125 - val_tp: 26.0000 - val_fp: 2.0000 - val_tn: 17.0000 - val_fn: 5.0000 - val_accuracy: 0.8600 - val_precision: 0.9286\n",
            "Epoch 2/20\n",
            "21/21 [==============================] - 9s 463ms/step - loss: 0.4440 - tp: 114.0000 - fp: 24.0000 - tn: 55.0000 - fn: 10.0000 - accuracy: 0.8325 - precision: 0.8261 - val_loss: 0.3361 - val_tp: 26.0000 - val_fp: 1.0000 - val_tn: 18.0000 - val_fn: 5.0000 - val_accuracy: 0.8800 - val_precision: 0.9630\n",
            "Epoch 3/20\n",
            "21/21 [==============================] - 9s 441ms/step - loss: 0.3521 - tp: 108.0000 - fp: 13.0000 - tn: 66.0000 - fn: 16.0000 - accuracy: 0.8571 - precision: 0.8926 - val_loss: 0.2644 - val_tp: 30.0000 - val_fp: 2.0000 - val_tn: 17.0000 - val_fn: 1.0000 - val_accuracy: 0.9400 - val_precision: 0.9375\n",
            "Epoch 4/20\n",
            "21/21 [==============================] - 9s 441ms/step - loss: 0.3380 - tp: 110.0000 - fp: 16.0000 - tn: 63.0000 - fn: 14.0000 - accuracy: 0.8522 - precision: 0.8730 - val_loss: 0.2540 - val_tp: 29.0000 - val_fp: 3.0000 - val_tn: 16.0000 - val_fn: 2.0000 - val_accuracy: 0.9000 - val_precision: 0.9062\n",
            "Epoch 5/20\n",
            "21/21 [==============================] - 10s 459ms/step - loss: 0.3134 - tp: 120.0000 - fp: 25.0000 - tn: 54.0000 - fn: 4.0000 - accuracy: 0.8571 - precision: 0.8276 - val_loss: 0.2485 - val_tp: 26.0000 - val_fp: 1.0000 - val_tn: 18.0000 - val_fn: 5.0000 - val_accuracy: 0.8800 - val_precision: 0.9630\n",
            "Epoch 6/20\n",
            "21/21 [==============================] - 10s 457ms/step - loss: 0.2888 - tp: 114.0000 - fp: 12.0000 - tn: 67.0000 - fn: 10.0000 - accuracy: 0.8916 - precision: 0.9048 - val_loss: 0.1708 - val_tp: 30.0000 - val_fp: 1.0000 - val_tn: 18.0000 - val_fn: 1.0000 - val_accuracy: 0.9600 - val_precision: 0.9677\n",
            "Epoch 7/20\n",
            "21/21 [==============================] - 11s 501ms/step - loss: 0.2573 - tp: 109.0000 - fp: 8.0000 - tn: 71.0000 - fn: 15.0000 - accuracy: 0.8867 - precision: 0.9316 - val_loss: 0.1767 - val_tp: 29.0000 - val_fp: 2.0000 - val_tn: 17.0000 - val_fn: 2.0000 - val_accuracy: 0.9200 - val_precision: 0.9355\n",
            "Epoch 8/20\n",
            "21/21 [==============================] - 10s 481ms/step - loss: 0.2401 - tp: 116.0000 - fp: 10.0000 - tn: 69.0000 - fn: 8.0000 - accuracy: 0.9113 - precision: 0.9206 - val_loss: 0.1965 - val_tp: 29.0000 - val_fp: 1.0000 - val_tn: 18.0000 - val_fn: 2.0000 - val_accuracy: 0.9400 - val_precision: 0.9667\n",
            "Epoch 9/20\n",
            "21/21 [==============================] - 9s 441ms/step - loss: 0.2534 - tp: 119.0000 - fp: 18.0000 - tn: 61.0000 - fn: 5.0000 - accuracy: 0.8867 - precision: 0.8686 - val_loss: 0.2267 - val_tp: 27.0000 - val_fp: 0.0000e+00 - val_tn: 19.0000 - val_fn: 4.0000 - val_accuracy: 0.9200 - val_precision: 1.0000\n",
            "Epoch 10/20\n",
            "21/21 [==============================] - 9s 446ms/step - loss: 0.2362 - tp: 117.0000 - fp: 11.0000 - tn: 68.0000 - fn: 7.0000 - accuracy: 0.9113 - precision: 0.9141 - val_loss: 0.1699 - val_tp: 28.0000 - val_fp: 0.0000e+00 - val_tn: 19.0000 - val_fn: 3.0000 - val_accuracy: 0.9400 - val_precision: 1.0000\n",
            "Epoch 11/20\n",
            "21/21 [==============================] - 9s 447ms/step - loss: 0.2177 - tp: 118.0000 - fp: 11.0000 - tn: 68.0000 - fn: 6.0000 - accuracy: 0.9163 - precision: 0.9147 - val_loss: 0.1346 - val_tp: 30.0000 - val_fp: 0.0000e+00 - val_tn: 19.0000 - val_fn: 1.0000 - val_accuracy: 0.9800 - val_precision: 1.0000\n",
            "Epoch 12/20\n",
            "21/21 [==============================] - 9s 444ms/step - loss: 0.2274 - tp: 118.0000 - fp: 8.0000 - tn: 71.0000 - fn: 6.0000 - accuracy: 0.9310 - precision: 0.9365 - val_loss: 0.1360 - val_tp: 29.0000 - val_fp: 0.0000e+00 - val_tn: 19.0000 - val_fn: 2.0000 - val_accuracy: 0.9600 - val_precision: 1.0000\n",
            "Epoch 13/20\n",
            "21/21 [==============================] - 9s 441ms/step - loss: 0.2057 - tp: 114.0000 - fp: 8.0000 - tn: 71.0000 - fn: 10.0000 - accuracy: 0.9113 - precision: 0.9344 - val_loss: 0.1549 - val_tp: 30.0000 - val_fp: 1.0000 - val_tn: 18.0000 - val_fn: 1.0000 - val_accuracy: 0.9600 - val_precision: 0.9677\n",
            "Epoch 14/20\n",
            "21/21 [==============================] - 9s 447ms/step - loss: 0.1577 - tp: 123.0000 - fp: 7.0000 - tn: 72.0000 - fn: 1.0000 - accuracy: 0.9606 - precision: 0.9462 - val_loss: 0.1650 - val_tp: 29.0000 - val_fp: 0.0000e+00 - val_tn: 19.0000 - val_fn: 2.0000 - val_accuracy: 0.9600 - val_precision: 1.0000\n",
            "Epoch 15/20\n",
            "21/21 [==============================] - 9s 443ms/step - loss: 0.1535 - tp: 119.0000 - fp: 6.0000 - tn: 73.0000 - fn: 5.0000 - accuracy: 0.9458 - precision: 0.9520 - val_loss: 0.1652 - val_tp: 27.0000 - val_fp: 1.0000 - val_tn: 18.0000 - val_fn: 4.0000 - val_accuracy: 0.9000 - val_precision: 0.9643\n",
            "Epoch 16/20\n",
            "21/21 [==============================] - 9s 443ms/step - loss: 0.1662 - tp: 119.0000 - fp: 8.0000 - tn: 71.0000 - fn: 5.0000 - accuracy: 0.9360 - precision: 0.9370 - val_loss: 0.1710 - val_tp: 28.0000 - val_fp: 0.0000e+00 - val_tn: 19.0000 - val_fn: 3.0000 - val_accuracy: 0.9400 - val_precision: 1.0000\n",
            "Epoch 17/20\n",
            "21/21 [==============================] - 9s 441ms/step - loss: 0.1461 - tp: 121.0000 - fp: 5.0000 - tn: 74.0000 - fn: 3.0000 - accuracy: 0.9606 - precision: 0.9603 - val_loss: 0.1563 - val_tp: 29.0000 - val_fp: 0.0000e+00 - val_tn: 19.0000 - val_fn: 2.0000 - val_accuracy: 0.9600 - val_precision: 1.0000\n",
            "Epoch 18/20\n",
            "21/21 [==============================] - 9s 440ms/step - loss: 0.1434 - tp: 121.0000 - fp: 3.0000 - tn: 76.0000 - fn: 3.0000 - accuracy: 0.9704 - precision: 0.9758 - val_loss: 0.1510 - val_tp: 29.0000 - val_fp: 1.0000 - val_tn: 18.0000 - val_fn: 2.0000 - val_accuracy: 0.9400 - val_precision: 0.9667\n",
            "Epoch 19/20\n",
            "21/21 [==============================] - 9s 441ms/step - loss: 0.1519 - tp: 118.0000 - fp: 5.0000 - tn: 74.0000 - fn: 6.0000 - accuracy: 0.9458 - precision: 0.9593 - val_loss: 0.1460 - val_tp: 29.0000 - val_fp: 0.0000e+00 - val_tn: 19.0000 - val_fn: 2.0000 - val_accuracy: 0.9600 - val_precision: 1.0000\n",
            "Epoch 20/20\n",
            "21/21 [==============================] - 9s 444ms/step - loss: 0.1231 - tp: 121.0000 - fp: 5.0000 - tn: 74.0000 - fn: 3.0000 - accuracy: 0.9606 - precision: 0.9603 - val_loss: 0.1502 - val_tp: 29.0000 - val_fp: 0.0000e+00 - val_tn: 19.0000 - val_fn: 2.0000 - val_accuracy: 0.9600 - val_precision: 1.0000\n"
          ]
        }
      ],
      "source": [
        "history2 = model.fit(train_set, epochs = 20, validation_data = valid_set)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 54,
      "metadata": {
        "id": "WysGt6bw03dB"
      },
      "outputs": [],
      "source": [
        "model.save('/content/brain_tumor_detection_model.h5')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 55,
      "metadata": {
        "id": "K6bPy-GW1Kum"
      },
      "outputs": [],
      "source": [
        "! cp /content/brain_tumor_detection_model.h5 /content/drive/MyDrive/"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 52,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "aMjKmW1p0tAS",
        "outputId": "dec5a39f-af51-411e-f2cd-05a31f85ddf9"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "loss :  0.15041548013687134\n",
            "tp :  30.0\n",
            "fp :  0.0\n",
            "tn :  19.0\n",
            "fn :  1.0\n",
            "accuracy :  0.9800000190734863\n",
            "precision :  1.0\n"
          ]
        }
      ],
      "source": [
        "validation_results = model.evaluate(valid_set, verbose=0)\n",
        "\n",
        "for name, value in zip(model.metrics_names, validation_results):\n",
        "  print(name, ': ', value)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 56,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 437
        },
        "id": "sYbPMgTS1Zed",
        "outputId": "3f5a2eea-13a4-4368-e3ce-e376df6c19b5"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Non-tumor detected (True Negatives):  18\n",
            "Non-tumor incorrectly detected (False Positives):  1\n",
            "Tumor missed (False Negatives):  2\n",
            "Tumor detected (True Positives):  29\n",
            "Total case:  31\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 360x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "validation_predictions = model.predict(valid_set)\n",
        "\n",
        "labels = valid_set.labels\n",
        "\n",
        "cm = confusion_matrix(labels, validation_predictions > 0.5)\n",
        "plt.figure(figsize=(5,5))\n",
        "sns.heatmap(cm, annot=True, fmt=\"d\")\n",
        "plt.title('Confusion matrix @{:.2f}'.format(0.5))\n",
        "plt.ylabel('Actual label')\n",
        "plt.xlabel('Predicted label')\n",
        "\n",
        "print('Non-tumor detected (True Negatives): ', cm[0][0])\n",
        "print('Non-tumor incorrectly detected (False Positives): ', cm[0][1])\n",
        "print('Tumor missed (False Negatives): ', cm[1][0])\n",
        "print('Tumor detected (True Positives): ', cm[1][1])\n",
        "print('Total case: ', np.sum(cm[1]))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iXRNrJZI1kp3"
      },
      "source": [
        "## Results"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 57,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "nBjpLo8g1hXt",
        "outputId": "0640607e-df3e-49bf-c921-56e83640da43"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "dict_keys(['loss', 'tp', 'fp', 'tn', 'fn', 'accuracy', 'precision', 'val_loss', 'val_tp', 'val_fp', 'val_tn', 'val_fn', 'val_accuracy', 'val_precision'])"
            ]
          },
          "execution_count": 57,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "h = history2.history\n",
        "h.keys()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 58,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 312
        },
        "id": "hg9ahqnh1txl",
        "outputId": "3030361d-3a7c-4b6f-cb38-fdbcf2286763"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<function matplotlib.pyplot.show>"
            ]
          },
          "execution_count": 58,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.plot(h['accuracy'])\n",
        "plt.plot(h['val_accuracy'], c='red')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('%')\n",
        "plt.title('acc vs val_acc')\n",
        "plt.show"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 59,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 312
        },
        "id": "V5E_kMdP1xDF",
        "outputId": "0c6052cc-e425-470b-855f-0158f978d1c9"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<function matplotlib.pyplot.show>"
            ]
          },
          "execution_count": 59,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.plot(h['accuracy'])\n",
        "plt.plot(h['loss'], c='red')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('%')\n",
        "plt.title('acc vs loss')\n",
        "plt.show"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [],
      "name": "DA_Project.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}