I have very much interested to work on compute vision problems like video surveillance, video analysis using webcams and at the same time have reasonable experience in that.
Right now I am also working this kind of project only, where we are recording the computer screen through webcam and running some models on the captured image. So for the sake of debugging the model to improve accuracy, for visualization and all we have developed lot of python utility scripts like your requirements. So it does not take much time to achieve your target.