Accurate Segmentation of Optic Disc And Cup from Multiple Pseudo-labels by Noise-aware Learning

Tengjin Weng,Yang Shen,Zhidong Zhao,Zhiming Cheng,Shuai Wang

DOI: https://doi.org/10.48550/arXiv.2311.18496

2024-03-15

Abstract:Optic disc and cup segmentation plays a crucial role in automating the screening and diagnosis of optic glaucoma. While data-driven convolutional neural networks (CNNs) show promise in this area, the inherent ambiguity of segmenting objects and background boundaries in the task of optic disc and cup segmentation leads to noisy annotations that impact model performance. To address this, we propose an innovative label-denoising method of Multiple Pseudo-labels Noise-aware Network (MPNN) for accurate optic disc and cup segmentation. Specifically, the Multiple Pseudo-labels Generation and Guided Denoising (MPGGD) module generates pseudo-labels by multiple different initialization networks trained on true labels, and the pixel-level consensus information extracted from these pseudo-labels guides to differentiate clean pixels from noisy pixels. The training framework of the MPNN is constructed by a teacher-student architecture to learn segmentation from clean pixels and noisy pixels. Particularly, such a framework adeptly leverages (i) reliable and fundamental insight from clean pixels and (ii) the supplementary knowledge within noisy pixels via multiple perturbation-based unsupervised consistency. Compared to other label-denoising methods, comprehensive experimental results on the RIGA dataset demonstrate our method's excellent performance. The code is available at <a class="link-external link-https" href="https://github.com/wwwtttjjj/MPNN" rel="external noopener nofollow">this https URL</a>

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the decline in model performance caused by label noise in the optic disc and optic cup segmentation tasks. Specifically, optic disc and optic cup segmentation play an important role in the automated screening and diagnosis of optic nerve head abnormalities such as glaucoma. However, although data - driven convolutional neural networks (CNNs) have shown potential in this field, due to the inherent ambiguity of the boundaries between the segmented objects and the background, there is noise in the annotation process, which affects the performance of the model.

To address this challenge, the paper proposes a method based on the Multiple Pseudo - labels Noise - aware Network (MPNN), aiming to improve the accuracy of optic disc and optic cup segmentation by generating multiple pseudo - labels and using the consensus information in these pseudo - labels to distinguish clean pixels from noisy pixels. This method is trained through a teacher - student architecture, where the student network learns by minimizing the segmentation loss on clean pixels and the consistency loss on noisy pixels, while the teacher network updates its parameters through exponential moving average (EMA).

The main contributions of the paper include:
1. Proposing MPNN, which is the first method to improve the accuracy of optic disc and optic cup segmentation from the perspective of label denoising.
2. Designing the Multiple Pseudo - label Generation and Guided Denoising (MPGGD) module, which can separate reliable and unreliable information, provide accurate prior information, and avoid the negative impact of label noise on the network.

Experimental results show that this method performs excellently on the RIGA dataset and significantly outperforms other label denoising methods.