Project II : Face Mask Detection
Aim of the Project
To build a Web App using Deep Convolutional Neural Networks to predict whether a person is Wearing a Mask or Not
Life Cycle of the Project
- Collected the dataset from Kaggle which contained 12,000 images in total for both the classes
- Used Data Augmentation to generate extra data from the existing dataset
- Applied Transfer Learning using Pre-trained VGG16 model and built a High Performance Custom CNN Model providing an Accuracy of 99.60 % on the Testing Dataset
- Saved the DL model with an h5 extension
- Built a Web App using Python at the Backend with Flask API and HTML, CSS & JS serving at the Frontend
- The Web App allows the User to upload an image, makes predictions on this image using the saved DL model and sends a Prediction text indicating whether the person in the image is Wearing a Mask or Not
Results from the Project
- If a person is Wearing a Mask
- If a person is Not wearing a Mask
Technologies Used
| Python | Tensorflow | Keras | CNN | VGG16 |
| Flask | Matplotlib | Numpy |
Performance of the Model
-
Training Accuracy = 95.82 %
-
Validation Accuracy = 97.71 %
-
Testing Accuracy = 99.60 %
-
Training Loss = 10.53 %
-
Validation Loss = 6.8 %
-
Testing Loss = 1.2 %