Overcoming the Challenges of Long-Tail Distribution in Nighttime Vehicle Detection
Houwang Zhang, Leanne Lai Hang Chan · IEEE Intelligent Systems · 2024
As a basic task of the intelligent transportation system, night-time vehicle detection is associated with many challenges. Existing methods usually ignore significant challenges arising from the imbalanced class distribution between vehicles, which always leads to poor detection for vehicles belonging to tail classes. By analyzing existing solutions for long-tail object detection and considering the complex and diverse characteristics of night-time traffic scenarios, we propose an enhanced detection approach based on anomaly detection. In addition, to tackle disturbance from complex lights, we re-construct the loss function for background proposals, thus allowing the detector to pay more attention to hard-classified proposals and to learn to distinguish vehicle lights from disturbed light resources. Comprehensive experiments prove that compared with generic approaches, our proposed method can effectively solve the problem of long-tail distribution in night-time vehicle detection and improve the robustness in complex environments.