Towards Robust Adaptive Object Detection under Noisy Annotations

TL;DR

This paper introduces a novel framework called NLTE for robust domain adaptive object detection that effectively handles noisy annotations in source datasets, significantly improving detection performance under label noise.

Contribution

It formulates noisy DAOD and proposes NLTE, a framework with three modules to mitigate the impact of noisy labels in domain adaptive object detection.

Findings

NLTE improves mAP by 8.4% under 60% corrupted annotations.

NLTE approaches the performance of models trained on clean data.

Thorough evaluation validates NLTE's effectiveness on benchmark datasets.

Abstract

Domain Adaptive Object Detection (DAOD) models a joint distribution of images and labels from an annotated source domain and learns a domain-invariant transformation to estimate the target labels with the given target domain images. Existing methods assume that the source domain labels are completely clean, yet large-scale datasets often contain error-prone annotations due to instance ambiguity, which may lead to a biased source distribution and severely degrade the performance of the domain adaptive detector de facto. In this paper, we represent the first effort to formulate noisy DAOD and propose a Noise Latent Transferability Exploration (NLTE) framework to address this issue. It is featured with 1) Potential Instance Mining (PIM), which leverages eligible proposals to recapture the miss-annotated instances from the background; 2) Morphable Graph Relation Module (MGRM), which models the adaptation feasibility and transition probability of noisy samples with relation matrices; 3) Entropy-Aware Gradient Reconcilement (EAGR), which incorporates the semantic information into the discrimination process and enforces the gradients provided by noisy and clean samples to be consistent towards learning domain-invariant representations. A thorough evaluation on benchmark DAOD datasets with noisy source annotations validates the effectiveness of NLTE. In particular, NLTE improves the mAP by 8.4\% under 60\% corrupted annotations and even approaches the ideal upper bound of training on a clean source dataset.