I’m not going to review the entire detailed process of training an object detector using Histogram of Oriented Gradients (yet), simply because each step can be fairly detailed. But I wanted to take a minute and detail the general algorithm for training an object detector using Histogram of Oriented Gradients. It goes a little something like this: