I am looking to have an API(preferably in Java/Python) which consume from a HLS(Http Live Streaming video) and use open-cv to detect an Logo(one positive image of object will be provided) from the stream. Same object can appear in a video multiple times. Every time a object is detected, it should log a message that object is detected with timestamp.
Also, detailed steps of how the training of a positive image is done to generate a model is required in addition to the working code.
I can provide more details like image and HLS later.