Hi. I'm working on a project that requires user validation with an image or a video from a web cam. Basically, I need to detect a person and then identify this person, and make sure the person is the same during some period of time.
Right now I'm working through the different solutions, like API based, or should I try and create everything from scratch? My main concerns are the data traffics (a lot will be uploaded to API) and the processing power to perform recognition and tracking.