Insert data as videos(Feed1, Feed2, Feed3) Same Names Press S to start the capture and Q to end the capture After capture press Spcaebar or enter

After capturing a frame, the system promptly uses a ROI selector to choose the target person, extracting a unique face encoding and an object descriptor. A color histogram is computed from the selected region using HSV color space. The histogram is compared against others using correlation, yielding a score between -1 and 1, with 1 indicating a perfect match. In subsequent frames, YOLO detects all persons present. For each detection, the system attempts face recognition by comparing face encodings. If no face is detected, it falls back on matching the object descriptor to the stored target descriptor, ensuring reliable identification.

Sometimes, as it is continously trying to eval, the position of target, the program doesnt end Press q in video feed or end process in terminal, else do force stop

Hope you like it :)