Objects do not disappear: Video object detection by single-frame object location anticipation

TL;DR

This paper introduces a video object detection method that leverages smooth motion to anticipate object locations from a static keyframe, improving accuracy and efficiency while reducing annotation costs across multiple datasets.

Contribution

The paper presents a novel approach that uses motion anticipation from a single keyframe to enhance detection accuracy, efficiency, and reduce annotation effort in video object detection.

Findings

Improved mean average precision over state-of-the-art methods.

Enhanced computational efficiency by processing fewer frames.

Reduced annotation costs by annotating only keyframes.

Abstract

Objects in videos are typically characterized by continuous smooth motion. We exploit continuous smooth motion in three ways. 1) Improved accuracy by using object motion as an additional source of supervision, which we obtain by anticipating object locations from a static keyframe. 2) Improved efficiency by only doing the expensive feature computations on a small subset of all frames. Because neighboring video frames are often redundant, we only compute features for a single static keyframe and predict object locations in subsequent frames. 3) Reduced annotation cost, where we only annotate the keyframe and use smooth pseudo-motion between keyframes. We demonstrate computational efficiency, annotation efficiency, and improved mean average precision compared to the state-of-the-art on four datasets: ImageNet VID, EPIC KITCHENS-55, YouTube-BoundingBoxes, and Waymo Open dataset. Our source code is available at https://github.com/L-KID/Videoobject-detection-by-location-anticipation.