I’m trying to run object re-identification between 2 cameras. In order to make that work, I need to somehow compute features for bounding boxes. Could you help me figure that out? For example, I would like to be able to take my bounding boxes, run them through a convolutional net (like ResNet) and get something like features from the last conv layer.
Or maybe there is a way to compute other kinds of features (SIFT or something like that)