Trackers gives you clean, modular re-implementations of leading multi-object tracking algorithms released under the permissive Apache 2.0 license. You combine them with any detection model you already use.
[CVPRW'24] SoccerNet Game State Reconstruction: End-to-End Athlete Tracking and Identification on a Minimap (CVPR24 - CVSports workshop)
SoccerNet@CVPR | 1st place solution for Ball Action Spotting Challenge 2023
SoccerNet@CVPR | 1st place solution for Camera Calibration Challenge 2023
Soccer pitch segmentation and camera calibration in two steps. Step 1, pixelwise segmentation of an broacast image of a soccer game into six different clases defined by the line markings. Step 2, a differential-rendering optimizer that tries to estimate camera parameters from such segementations.Trained on SoccerNet.
Spiideo SoccerNet SynLoc - Single Frame World Coordinate Athlete Detection and Localization with Synthetic Data
A Baseline for Multi Objective Tracking (MOT) of Soccer and Soccer Players Based on SoccerNet Tracking Dataset and PaddleDetection.
[WINNER SOLUTION] soccernet monocular depth estimation solution
Robust Player Tracking & Behavior Analysis pipeline for SoccerNet Benchmark. Features YOLO11x, BoT-SORT (GMC), and Adaptive Field Masking. Artificial Vision Project 2025/26 @ Unisa.
Football match analysis on SoccerNet: player and ball detection, tracking, team assignment and pitch homography onto a tactical minimap.
Soccer video understanding — YOLO11m fine-tuned on SoccerNet, Kalman + ReID multi-object tracking, and ROI-based behavior analysis (Artificial Vision course project)
Football match analysis on SoccerNet: detection, tracking, team assignment, pitch calibration and tactical projection.
Predicting a referee's post-foul decision from player pose alone — 11,674 SoccerNet fouls, 17 COCO keypoints, 2-layer LSTM optimised for macro-F1 over accuracy.
SoccerNet 2026 VQA를 위한 SoccerAgent 도구 개선 프로젝트: 영상 검색·인원 계수·카메라 시점 분류
👤 Conditional diffusion model for 64x64 face generation on CelebA, conditioned on gender, smile, and age attributes. Implements a DDPM-based architecture for controllable face synthesis and attribute-guided sampling.