Seq-NMS for Video Object Detection

TL;DR

Seq-NMS enhances video object detection by leveraging high-scoring detections from neighboring frames to improve weaker detections, leading to superior results compared to single-image methods and achieving top placement in ILSVRC2015.

Contribution

The paper introduces Seq-NMS, a novel post-processing method that improves video object detection by utilizing temporal information from adjacent frames.

Findings

Outperforms state-of-the-art single-image detection methods

Achieved 3rd place in ILSVRC2015 VID task

Demonstrates the effectiveness of temporal score boosting

Abstract

Video object detection is challenging because objects that are easily detected in one frame may be difficult to detect in another frame within the same clip. Recently, there have been major advances for doing object detection in a single image. These methods typically contain three phases: (i) object proposal generation (ii) object classification and (iii) post-processing. We propose a modification of the post-processing phase that uses high-scoring object detections from nearby frames to boost scores of weaker detections within the same clip. We show that our method obtains superior results to state-of-the-art single image object detection techniques. Our method placed 3rd in the video object detection (VID) task of the ImageNet Large Scale Visual Recognition Challenge 2015 (ILSVRC2015).