Towards High Performance Video Object Detection
This work addresses video object detection, an important practical scenario, but it is incremental as it builds upon prior works with new techniques.
The paper tackles the problem of video object detection, which is more challenging than image detection, by proposing a unified approach based on multi-frame end-to-end learning and cross-frame motion, achieving steady improvements in the speed-accuracy tradeoff.
There has been significant progresses for image object detection in recent years. Nevertheless, video object detection has received little attention, although it is more challenging and more important in practical scenarios. Built upon the recent works, this work proposes a unified approach based on the principle of multi-frame end-to-end learning of features and cross-frame motion. Our approach extends prior works with three new techniques and steadily pushes forward the performance envelope (speed-accuracy tradeoff), towards high performance video object detection.