An EffcientNet-encoder U-Net Joint Residual Refinement Module with Tversky–Kahneman Baroni–Urbani–Buser loss for biomedical image Segmentation. (May 2023)

• Record Type: Journal Article
• Authors: Nham, Do-Hai-Ninh; Trinh, Minh-Nhat; Nguyen, Viet-Dung; Pham, Van-Truong; Tran, Thi-Thao
• Is Part Of: Biomedical signal processing and control. Volume 83(2023)
• Publication Date: 2023-05
• Subjects: Biomedical Image Segmentation -- EfficientNet -- Residual Refinement Module -- Baroni–Urbani–Buser coefficient -- Tversky–Kahneman probability-weighting function; 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.2023.104631
• Languages: English
• ISSNs: 1746-8094

Abstract

Abstract: Quantitative analysis on biomedical images has been on increasing demand nowadays and for modern computer vision approaches. While recently advanced procedures have been enforced, there is still necessity in optimizing network architecture and loss functions. Inspired by the pretrained EfficientNet-B4 and the refinement module in boundary-aware problems, we propose a new two-stage network which is called EffcientNet-encoder U-Net Joint Residual Refinement Module and we create a novel loss function called the Tversky–Kahneman Baroni–Urbani–Buser loss function. The loss function is built on the basement of the Baroni–Urbani–Buser coefficient and the Jaccard–Tanimoto coefficient and reformulated in the Tversky–Kahneman probability-weighting function. We have evaluated our algorithm on the four popular datasets: the 2018 Data Science Bowl Cell Nucleus Segmentation dataset, the Brain Tumor LGG Segmentation dataset, the Skin Lesion ISIC 2018 dataset and the MRI cardiac ACDC dataset. Several comparisons have proved that our proposed approach is noticeably promising and some of the segmentation results provide new state-of-the-art results. The code is available at https://github.com/tswizzle141/An-EffcientNet-encoder-U-Net-Joint-Residual-Refinement-Module-with-TK-BUB-Loss . Highlights: This paper presents a deep learning-based method for automatic biomedical image segmentation . The proposed EffcientNet-encoder U-Net Joint Residual Refinement Module includes two primary stages: U-Net with EfficientNet-B4 encoder and Residual Refinement Module. A novel loss function is proposed based on the Baroni–Urbani–Buser and the Jaccard–Tanimoto similarity coefficient and the reformulation in the Tversky–Kahneman probability-weighting function. The proposed approach obtains promising results and sets some new state-of-the-arts.