This project detects head rotations (left, right, up, down) in videos using YOLOv3 for person detection and Haar Cascade for face detection. The backend is built with Flask, and the frontend is a simple HTML interface for uploading videos and viewing detected frames.
Features Upload and process videos.
Detect persons using YOLOv3.
Detect head rotation using Haar Cascade.
Display detected frames with bounding boxes and head direction.
Setup Clone the repository.
Install dependencies: pip install flask opencv-python. Yolov3 weights:https://pjreddie.com/media/files/yolov3.weights Yolov3 configuration:https://github.com/pjreddie/darknet/blob/master/cfg/yolov3.cfg Run the app: python app.py.
Open http://localhost:5000 in your browser.
Usage Upload a video.
View detected frames with head rotation directions.