Score (sn44) is a decentralized computer vision and artificial intelligence project operating as Subnet 44 on the Bittensor network. The primary focus of the project is the collection, annotation, and real-time analysis of sports data, with an initial emphasis on global football. By leveraging the decentralized infrastructure of the Bittensor ecosystem, Score aims to create a scalable computer vision platform that provides deep insights into player performance and game dynamics. The technical core of the project involves using advanced machine learning models to perform object detection and keypoint analysis on video footage. This process allows the system to track individual player movements, ball positions, and specific game events in real time. The network functions through a competitive incentive structure where miners act as data annotators, processing video frames to identify key elements. Validators then evaluate the accuracy and speed of these annotations against ground truth data or actual match outcomes. The insights generated by the system are designed for a variety of users, including professional football clubs, scouts, sports bettors, and fantasy sports platforms. The project developed a football value function that assesses a player's impact on a game based on their real-time actions and positions. This data-driven approach is intended to provide tactical insights and predictive modeling that can outperform traditional centralized analytics providers. While sports analytics is the current priority, the project has a broader vision of becoming a universal vision AI layer. The underlying technology is being developed to handle any type of video footage, which would allow it to expand into industries such as security, retail, and logistics. A notable real-world application of this expansion is a partnership with a major European petrol station network to monitor operational efficiency and identify equipment failures using the same computer vision technology. The Score team consists of professionals with backgrounds in sports technology, web3, and machine learning. The project currently processes data from hundreds of football leagues worldwide, providing a high level of coverage for global matches. The SN44 token serves as the utility and incentive asset within this specific subnet, facilitating the decentralized coordination of compute resources and data verification. Through this model, Score attempts to democratize access to high-quality, automated sports data and advanced computer vision tools.
Learn more
