PNT-Edge: Towards Robust Edge Detection with Noisy Labels by Learning Pixel-level Noise Transitions

TL;DR

PNT-Edge introduces a novel approach to improve edge detection accuracy by modeling pixel-level label noise through noise transition learning and a shift estimation module, effectively handling noisy labels in large datasets.

Contribution

The paper proposes Pixel-wise Shift Learning to estimate noise transitions and a regularization term for better noise modeling in edge detection, addressing label noise issues.

Findings

Effective noise transition modeling improves edge detection accuracy.

Significant reduction in label noise impact demonstrated on SBD and Cityscapes datasets.

The method outperforms existing approaches in noisy label scenarios.

Abstract

Relying on large-scale training data with pixel-level labels, previous edge detection methods have achieved high performance. However, it is hard to manually label edges accurately, especially for large datasets, and thus the datasets inevitably contain noisy labels. This label-noise issue has been studied extensively for classification, while still remaining under-explored for edge detection. To address the label-noise issue for edge detection, this paper proposes to learn Pixel-level NoiseTransitions to model the label-corruption process. To achieve it, we develop a novel Pixel-wise Shift Learning (PSL) module to estimate the transition from clean to noisy labels as a displacement field. Exploiting the estimated noise transitions, our model, named PNT-Edge, is able to fit the prediction to clean labels. In addition, a local edge density regularization term is devised to exploit local structure information for better transition learning. This term encourages learning large shifts for the edges with complex local structures. Experiments on SBD and Cityscapes demonstrate the effectiveness of our method in relieving the impact of label noise. Codes are available at https://github.com/DREAMXFAR/PNT-Edge.