MatchIQ: AI-Powered Football Match Analytics Platform

Loading...

Media is loading
 

Degree

Bachelor of Science (Computer Science)

Department

Department of Computer Science

School

School of Mathematics and Computer Science (SMCS)

Advisor

Dr. Imran Rauf, Assistant Professor, Department of Computer Science

Keywords

Football Analytics, Computer Vision, Object Detection, Multi-Object Tracking, Player Performance Analysis, YOLOv8, Sports Analytics

Abstract

MatchIQ is an AI-powered football analytics system designed to automate player tracking, performance analysis, and match insight generation using computer vision and machine learning techniques. The system processes football match recordings to detect and track players, referees, goalkeepers, and the ball in real time or near real time. The project utilizes the YOLOv8 object detection model for accurate entity detection and the HybridSort tracking algorithm to maintain player identities across video frames. Using the extracted tracking information, the system computes key performance metrics such as player speed, acceleration, movement trajectories, total distance covered, fatigue estimation, ball possession statistics, and shot analysis. In addition, MatchIQ integrates predictive analytics techniques to estimate goal probability and predict match outcomes using historical and extracted match features. The generated insights are presented through an interactive dashboard with graphical visualizations, enabling coaches, analysts, and teams to better understand player and team performance. By automating football match analysis, MatchIQ reduces the need for manual data collection and provides an efficient, scalable, and data-driven solution for modern sports analytics.

Tools and Technologies Used

Python, YOLOv8, HybridSort, OpenCV, FastAPI, Next.js, WebSocket, KMeans Clustering, OSNet Re-ID, Homography Transformation, MJPEG Streaming, CSV, JavaScript.

Methodology

MatchIQ follows a modular pipeline-based development approach. The system begins with video input, which is decomposed into frames for frame-level analysis. YOLOv8 is used for object detection to identify players, referees, goalkeepers, and the ball. The HybridSort tracking algorithm is then applied to maintain consistent player identities across frames. Tracking outputs are used to extract movement-based features including speed, acceleration, distance covered, and trajectories using mathematical frame-to-frame positional modeling. KMeans clustering is applied for team classification, and homography transformation is used for pitch mapping and spatial analysis. Higher-level analytics including fatigue estimation, goal probability, and match outcome prediction are computed using EMA-based heuristics and weighted linear models. All results are served through a FastAPI backend via REST APIs and WebSocket connections, and visualized through an interactive Next.js dashboard with real-time MJPEG streaming and a PitchRadar minimap.

Document Type

Restricted Access

Submission Type

BSCS Final Year Project

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

This document is currently not available here.

Share

COinS