The person-detection was a prototype dnn model to detect person from a live feed. The pretrained darknet neural network is implemented in the model. The darknet could detect 81 objects out of which person detection was only implemented. In a real time feed person will be detected and a rectangular box will be depict them. The program will detect persons in the real time images and count the persons present in the frame and save that count to database. Database used was mongodb.