Since CCTVs have the high ground ( ͡° ͜ʖ ͡°) simply comparing the distances between the centroids should tell us if people are social distancing or not.
- Implement some object tracking method. Tracking multiple objects using KCF is VERY slow.
- Clean the code.
- Train a better model for this use-case.
... will add more when I think of them
yolov3.cfg - https://github.com/pjreddie/darknet/blob/master/cfg/yolov3.cfg
yolov3.weights - https://pjreddie.com/media/files/yolov3.weights
Tensorflow MobileNet-SSD v1 - http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v1_coco_2017_11_17.tar.gz and https://github.com/opencv/opencv_extra/blob/master/testdata/dnn/ssd_mobilenet_v1_coco_2017_11_17.pbtxt