README.md

Deployment_PROJECT (Face Mask Detector) Using Streamlit and Heroku by Applying Pre- Trained CNN Model (MobileNetV2)

by Mohamed Sebaie Sebaie

This is a simple Streamlit frontend for face mask detection in images using a pre-trained Keras CNN model MobileNetV2 and OpenCV then deploy on heroku.

The Web Application I Created, is in This Link.

The Data used for training can be found through This Link on Kaggle Website.

All work here is done on CoLab

General Info

• This Project has been implemented by using OpenCV to detect faces in the input images and a a pre-trained Keras CNN model (MobileNetV2) as mask/no-mask binary classifier applied to the faces Images. The Deep Learning model currently used has been trained using this image data set from kaggle here . The trained model has been shared in this repo. The face detector algorithm comes from here: the Caffee model files are in CAFFEE folder directory.

Web APP Explanation

Once an image has been uploaded, the classification happens automatically.

About The Data:

The dataset used for Training consists of one zip file Face Mask Dataset that is download in Colab and unzipped then Create a pre-trained Keras CNN model (MobileNetV2) and Training then evaluate, save and test the model. The NoteBooks are in face_mask_detector_notebooks.

Finally, After creating the Model and save as h5 file, Deploy the model with Streamlit frontend and upload it toHeroku Platform..

The Web Application I Created, is in This Link.

Good Reference for Deployment a Streamlit Frontend to Heroku here.