Unsupervised Deep learning to track fish

I am looking for a freelancer to use unsupervised deep learning and AI to track zebrafish using 2D tracking from video data. There are a lot of challenges associated with tracking these fish, and I'm looking for someone with the necessary skills to develop a reliable tracking model, using methods such as image segmentation and object recognition. The ultimate goal is to have an accurate model that can detect, track, and classify different fish in real-time, even when they appear in groups. The model should be able to differentiate between fish, so that the fish can be identified and tracked throughout the duration of the video. Furthermore, the model should be able to detect and quantify any behavioural changes or motion evolutions in the fish. If possible, I would like the model to be integrated with an existing fish tracking system, in order to improve its accuracy and precision. This project will require expertise in computer vision and machine learning, as well as a keen eye for detail. The final product should be a functioning system, capable of accurately tracking zebrafish in real-time.