Shawn Abraham (URK20CS3026)

Summer Intern

Intel Corporation•  May 2023 - July 2023

The initial week of my internship was the orientation session. In the following weeks, Technical and practice sessions were conducted by Industry mentors which helped us learn about the use of OPENVINO toolkit optimizations for deep learning models. This was done in parallel with the project. Social distance Monitoring using Computer Vision and deep learning. I made use of OpenCV for loading either image or video or camera feed to be processed, and did some preprocessing like colour correction and resizing to obtain input frame. The inputs are fed into a YOLO v8 model, which has been fine-tuned for person detection. I have done some other optimisations of its parameters. It gave an output tensor containing the coordinates of the bounding boxes of detected people. I further made a dynamic distance measuring algorithm (on the basis of the location of the person in the frame) cum Social Distancing Violation counting function which returned the number and id(w.r.t to detections) of violators. Using the detections bounding boxes tensor and information on the id of violators, they were highlighted on the output frame in a red rectangle with a line connecting every pair of social distance violations and displayed the number of violations in a corner of the screen. I also made use of the OPENVINO toolkit.