YOLOv7 based face extraction and novel Eye feature extraction with EYENET

Most of the modern Face feature techniques include face detection and extraction. The novel techniques include feature extraction from the face which are mapped with the names of person used for detecting the person on a live basis. The project starts with a YOLO (You look only once) v7 model which was the high accuracy and prediction model created until date. This model is basically used for object detection, has also been used for face detection purposes. The model is further combined with computer vision color mapping methods to extract the face out. Then a feature-based model is built to extract 12 feature points in eyes from the face. Every person is identified according to the values like length and area from the feature points of the eyes. Comparison between the YOLO v7 and other face detection and feature extraction methods have also been specified. The main aim of the paper is for extraction of certain features and few novel feature techniques that can be used in person re-identification.