README:

Final Report:

https://sheldonsebastian.github.io/face_mask_detector/

Directory Structure:

path	Description
Code/Data Download Scripts	Folder containing scripts to download the images from one drive and google drive using CURL.
Code/DataPreprocessing.py	To split the downloaded images into train-validation-holdout and organize the files for ImageFolder.
Code/Train.py	Training of the final resnet50 model.
Code/Inference.py	To view the inference probability of each class based on input image passed.
Code/OcclusionExperiment.py	To intrepret and visualize the CNN based on occlusion experiment
docs/	Files related to github website
saved_model	Folder containing the trained model
Final-Group-Presentation.pdf	presentation slides as pdf
Final-Group-Project-Report.pdf	final report as pdf

Steps to replicate project:

1. Download data using Data Download Scripts
2. Train model using Train.py
3. Perform Inference using Inference.py
4. To interpret the model use OcclusionExperiment.py