Description
Computer-vision projects can detect, track, segment, classify, and estimate poses in images or video. This toolkit provides Ultralytics model workflows for training, inference, evaluation, and export from Python or its command-line interface.
It can process local media, datasets, streams, or camera-like inputs depending on how it is used. Treat faces, locations, and other visual data as privacy-sensitive, and validate model results before relying on automated decisions.