Рет қаралды 12,264
In this live session we'll use the Football Player Detection dataset[1] by Scott Bronkema[2] shared on Roboflow Universe to create a program that tracks players' movements using zero-shot object detection[3] in a football play and visualizes the result.
Tools we'll use/cover include: Python, Roboflow, YOLOv4, YOLOv5, CLIP, zero shot object tracking, ffmpeg, and OpenCV.
Update: we sent an improved dataset to training at the end of the video; we'll pick up here once the improved model is finished training with part 2. Subscribe and ding the bell to be notified!
[1] universe.robof...
[2] / scott-bronkema-09389824
[3] blog.roboflow....