Sequence Level Semantics Aggregation for Video Object Detection

TL;DR

This paper introduces a novel sequence-level semantics aggregation (SELSA) module for video object detection, improving robustness and discriminability of features by aggregating information across entire video sequences, leading to state-of-the-art results.

Contribution

The paper proposes a new sequence-level feature aggregation method, SELSA, that enhances video object detection without relying on optical flow or recurrent networks.

Findings

Achieves new state-of-the-art results on ImageNet VID and EPIC KITCHENS datasets.

Does not require complex postprocessing like Seq-NMS or Tubelet rescoring.

Demonstrates the connection between SELSA and spectral clustering methods.

Abstract

Video objection detection (VID) has been a rising research direction in recent years. A central issue of VID is the appearance degradation of video frames caused by fast motion. This problem is essentially ill-posed for a single frame. Therefore, aggregating features from other frames becomes a natural choice. Existing methods rely heavily on optical flow or recurrent neural networks for feature aggregation. However, these methods emphasize more on the temporally nearby frames. In this work, we argue that aggregating features in the full-sequence level will lead to more discriminative and robust features for video object detection. To achieve this goal, we devise a novel Sequence Level Semantics Aggregation (SELSA) module. We further demonstrate the close relationship between the proposed method and the classic spectral clustering method, providing a novel view for understanding the VID problem. We test the proposed method on the ImageNet VID and the EPIC KITCHENS dataset and achieve new state-of-the-art results. Our method does not need complicated postprocessing methods such as Seq-NMS or Tubelet rescoring, which keeps the pipeline simple and clean.