Abdurashid

Tactical Twin Vision is a football video-analysis prototype powered by Meta SAM 3.1. Users upload football footage, and the application processes the video using Meta's segmentation and object-tracking API. It visualizes detected players with animated bounding boxes, segmentation masks, persistent track IDs, and frame-synchronized playback. Beyond the standard model output, Tactical Twin Vision provides interactive tracking timelines, movement trajectories, player-count statistics, spatial observations, and evidence of tracking gaps. Built with Python, FastAPI, React, and Meta Model API, the project demonstrates how computer vision can support more accessible football analysis. Current limitations include occasional missed detections, lack of verified team identification, and no pitch-calibrated tactical conclusions.