Face-detection-using-vgg16

Project Overview: Face Detection using VGG16 and TensorFlow

Files:

1. face_detection.ipynb: Contains the model training with 99% validation accuracy.
2. live_face_detection.py: Utilizes OpenCV for real-time face detection.
3. Demo Images: For testing purposes.
4. screen_shot.png (1 & 2): Captures of live face detection.
5. hasrcasecade_face_frontage_default.xml: Used for detecting face shapes in live footage.

Dependencies:

1. Tensorflow(1.13.1)
2. Keras(2.3.1)
3. OpenCV2(4.1.1)
4. Python(3.7.7)
5. Anaconda(4.7.11)

Setup Instructions:

1. Download the entire repository.
2. Use your dataset.
3. Open and run all cells in face_detection.ipynb using Jupyter Notebook.

Notebook Workflow:

• Import necessary libraries.
• Define variables (e.g., image_size, train/test paths).
• Download and integrate VGG16.
• Construct and train the model, validate, and plot accuracies.
• Test the model using a sample image.
• Save the model as 'Final_Model_Face.h5'.

Live Detection (live_face_detection.py):

• Import libraries and load 'Final_Model_Face.h5'.
• Ensure webcam functionality.
• Use 'haarcascade_frontalface_default.xml' for live facial detection.
• Convert detected faces to array for model prediction.
• Display predictions (Ash, Malav, Nani) on live camera feed.
• For any issues, check paths and installed modules.

Thank you.
Ashish
linkedin : https://www.linkedin.com/in/ashishbarvaliya/