R&D 100 winner of the day: Traffic Violation Catcher

In Taiwan, traffic violations, especially those involving motorcycles, remain a problem. In 2022 alone, traffic accidents resulted in 3,085 deaths and nearly 500,000 injuries, the highest in a decade. Limited police resources make traffic violation enforcement difficult, leading to a gap between issued penalties and actual infractions. Traditional enforcement methods pose safety risks to officers, with an average of one traffic officer injured in a traffic accident every nine days. The Institute for Information Industry (III) has developed AI-based recognition technology to detect real-time traffic violations and address these issues. The “Traffic Violation Catcher” application uses sensors and deep learning to automatically identify violations such as running red lights and failing to yield to pedestrians. Compared to traditional methods, AI-powered traffic enforcement reduces officers’ workload, minimizes their exposure to danger, and ensures fairness. The system operates 24/7, providing concrete evidence and facilitating penalties.