FaceID | Yolo | Dataset Generation

15BCE126 – Yash Thesia 15BCE127 – Tirth Pandya

F.Y.D or Face ID, Yolo and Dataset Generation is a fusion of next generation Computer Vision and Training paradigms. It is an innovative Web Application that registers users and enables them to upload photos onto a server database. After collecting a fair amount of images from different users, the application runs different Object Detection and Face Recognition algorithms ( Yaar Cascades | Yolo ) in the background. vThe Result of the detection and recognition are then displayed to the user, either as a static image or as a Real-Time Video.

The project is a fusion of three major components :

1. Dataset Generation.
2. Facial Recognition using Haar Cascades.
3. Object Detection using YOLO.

Our objective was to first be able to collect a generous amount of images as dataset in the server. Secondly, we had to apply the algorithms of Facial Feature extraction included in the Haar Cascades library of OpenCV. Next, we trained the images using the Trainer Function, and then the Recognizer Function was used to Tell who the user is in real-time. Finally, we enhanced both the components : The Web UI for Dataset Collection, and the Python Script for detection and recognition and fused them into a single application : The F.Y.D.

numpy, pandas, matplotlib, tensorflow, keras, scikit-learn, Darknet, and other built-in libraries.

Yash_Thesia_FaceID_Report.pdf - contains datail discription of the exectution of the code and its expected results.

Please refer to inline comments in python scripts for detailed descriptions.