Dense Learning based Semi-Supervised Object Detection

TL;DR

This paper introduces a novel dense learning approach for semi-supervised object detection that is specifically designed for anchor-free detectors, achieving state-of-the-art results on standard benchmarks.

Contribution

It proposes a new anchor-free SSOD algorithm with techniques like adaptive filtering, an aggregated teacher, and uncertainty regularization to improve detection accuracy.

Findings

Achieves new state-of-the-art SSOD performance on MS-COCO and PASCAL-VOC.

Surpasses existing methods by a large margin.

Demonstrates effectiveness of dense pixel-wise pseudo-labeling and uncertainty regularization.

Abstract

Semi-supervised object detection (SSOD) aims to facilitate the training and deployment of object detectors with the help of a large amount of unlabeled data. Though various self-training based and consistency-regularization based SSOD methods have been proposed, most of them are anchor-based detectors, ignoring the fact that in many real-world applications anchor-free detectors are more demanded. In this paper, we intend to bridge this gap and propose a DenSe Learning (DSL) based anchor-free SSOD algorithm. Specifically, we achieve this goal by introducing several novel techniques, including an Adaptive Filtering strategy for assigning multi-level and accurate dense pixel-wise pseudo-labels, an Aggregated Teacher for producing stable and precise pseudo-labels, and an uncertainty-consistency-regularization term among scales and shuffled patches for improving the generalization capability of the detector. Extensive experiments are conducted on MS-COCO and PASCAL-VOC, and the results show that our proposed DSL method records new state-of-the-art SSOD performance, surpassing existing methods by a large margin. Codes can be found at \textcolor{blue}{https://github.com/chenbinghui1/DSL}.