Towards High Performance Video Object Detection

TL;DR

This paper introduces a unified multi-frame end-to-end learning approach for video object detection, incorporating three new techniques to improve speed and accuracy in practical scenarios.

Contribution

It extends prior methods with three novel techniques, advancing the performance of high-quality video object detection.

Findings

Achieved improved speed-accuracy tradeoff in video detection

Demonstrated effectiveness of multi-frame end-to-end learning

Pushed forward the performance envelope in practical scenarios

Abstract

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.