FaceMaskDetection

• Repo contains two .ipynb notebooks

  1.Face_Mask_detector_for_Covid_19 :

  Notebook contains Code for Data Preprocessing and Training using Face Detector Model built from scratch using Keras

  2.Face_mask_detection_using_MobileNetV2 :

  Notebook Contains Code for Data Preprocessing and Training using Finetuned MobileNetV2 model pretrained on imagenet dataset for face mask detection

Face Mask Detector is built using multi-step approach by extracting face of person using HaarCascades & passing it to mask detection model

Results and Accuracy of Face_Mask_detector_for_Covid_19 Notebook is as below:

Architecture of mask detection model

Loss and Accuracy (F Score) of model

Results

1.Results on test data

2.Results on Random Images from google

3.If Face is not detected by Haar Cascades, whole image is fed to model and model performs fine on it

Results and Accuracy of Face_mask_detection_using_MobileNetV2 Notebook is as below:

Loss and Accuracy (F Score) of model

Results