Skip to content

Detecting Skin Cancer Melanoma (malignant melanoma) using Deep Learning based methods

Notifications You must be signed in to change notification settings

tkrsh/DetectMelanoma

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

23 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Detecting Skin Cancer Melanoma (malignant melanoma)

Melanoma occurs when the pigment-producing cells that give colour to the skin become cancerous. Symptoms might include a new, unusual growth or a change in an existing mole. Melanomas can occur anywhere on the body. Treatment may involve surgery, radiation, medication or in some cases, chemotherapy.

Dataset

Society for Imaging Informatics in Medicine

SIIM-ISIC Melanoma Classification

The images were provided in DICOM format JPEG and TFRecord format.

Download Dataset

Hardware Used For Training:

  1. TPUs on Kaggle
  2. Local Machine Nvidia GPU GTX 1660 Ti
  3. Inference done on CPU

Library and Language Used :

Python 3.6

  1. TensorFlow I/0 // DICOM handling
  2. TensorFlow 2.3 // Deep Learning Model Implementation
  3. Pydicom 2.0 // DICOM handling
  4. OpenCv 3.2 // Image Preprocessing
  5. Pandas 1.13 // Csv Handling

Conv Models Used :

  1. EfficientNetb0-b7
  2. DenseNet 169

Preprocessing Used (Image):

  1. Resizing , zoom and croping
  2. Image Augmentation
1) Rotation_range = 180
2) shear_range = 0.4
3) Horizontal And Vertical Flipping 
4) Rescale 

Other Techniques Used:

Generated patient metadadata from DICOM

Performed 2D/3D layering utilizing different windows and slicing of DICOM Image

Result Obtained :

ROC = 0.93

Submissions were evaluated on area under the ROC curve between the predicted probability and the observed target.

About

Detecting Skin Cancer Melanoma (malignant melanoma) using Deep Learning based methods

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published