Intelligent Match Counter in a Video Game Arena

Abstract

This paper presents the development of an intelligent match counter for video game arenas, a rapidly expanding sector where manual match counting is pivotal but prone to errors, often resulting in financial losses. The proposed solution introduces an automated system that leverages cutting-edge advancements in computer vision, including convolutional neural networks (CNNs) for scoreboard detection and optical character recognition (OCR) for text extraction. The system accurately extracts pertinent data, such as team names, scores, and game time, and performs match counting based on data variations. A comprehensive methodology was adopted, combining functional and structural modeling, hardware integration utilizing a Raspberry Pi nano-computer, and the design of an ergonomic, 3D-printed enclosure. Real-world testing confirms the system’s efficacy, meeting the precision and real-time processing demands of the arena environment. This work represents a significant advancement in the modernization of video game arena management, improving transparency, profitability, and operational efficiency. Moreover, it offers a scalable solution that can be adapted to other automated management applications.