Exploiting Temporal Context for Tiny Object Detection

In surveillance applications, the detection of tiny, lowresolution objects remains a challenging task. Most deep learning object detection methods rely on appearance features extracted from still images and struggle to accurately detect tiny objects. In this paper, we address the problem of tiny object detection for real-time surveillance...