Multi-rater label fusion based on an information bottleneck for fundus image segmentation. (January 2023)

• Record Type: Journal Article
• Authors: Zhang, Feiyan; Zheng, Yuanjie; Wu, Jie; Yang, Xinbo; Che, Xiaowei
• Is Part Of: Biomedical signal processing and control. Volume 79(2023)Part 1
• Publication Date: 2023-01
• Subjects: Multi-rater annotation -- Information bottleneck -- Fundus image segmentation; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2022.104108
• Languages: English
• ISSNs: 1746-8094

Abstract

Abstract: In the fundus image segmentation process, more than one professional rater is usually required to mark the organ in the image to reduce the error rate of diagnosis. However, it is undeniable that each rater has an independent level of expertise and experience, and there are often individual differences in annotation areas, which can produce redundant or irrelevant information. In computer vision, for feature selection of multiple annotations, the majority voting principle or the preferred annotator principle is usually used, which cannot effectively remove the noise inside the annotator and the redundant information between annotators. To solve the above problems, we studied the information bottleneck method to remove the annotation noise and extend it to multi-expert annotation tasks in the fundus image field to extract consistent information between different views. We believe that consistent information is the key information that multiple experts agree on and have directivity. To the best of our knowledge, this is the first model for extracting multi-expert consistency information via multi-view information bottlenecks. Specifically, we use a multi-view information bottleneck approach to obtain the most concise expression of each view under label supervision. In addition, we propose a novel unsupervised information bottleneck method by maximizing mutual information between multiple view representations to preserve consistent information while eliminating redundant information that is not shared between views. A large number of experiments on several public datasets prove the effectiveness of the proposed model, and its performance is superior to existing techniques. Highlights: We introduced a fundus image segmentation model MRIBNet that fuses annotated labels from multiple experts and segments. We theoretically extend the Information Bottleneck(IB) to multi-rater IB in an unsupervised setting, which can extract the most consistent information from multi-rater annotations. The proposed method can be viewed as a novel label fusion strategy focusing on extracting highly correlated features between labels. The proposed method achieves the high accuracy on both public and private datasets of fundus images.