DOLG-NeXt: Convolutional neural network with deep orthogonal fusion of local and global features for biomedical image segmentation

被引:6
作者
Ahmed, Md. Rayhan [1 ]
Fahim, Asif Iqbal [1 ]
Islam, A. K. M. Muzahidul [1 ]
Islam, Salekul [1 ]
Shatabda, Swakkhar [1 ]
机构
[1] United Int Univ, Dept Comp Sci & Engn, Plot-2, United City, Madani Ave, Dhaka 1212, Bangladesh
关键词
Multi-scale information aggregation; Biomedical image segmentation; ConvNeXt; Deep orthogonal fusion of local and global features; Squeeze and excitation networks;
D O I
10.1016/j.neucom.2023.126362
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Biomedical image segmentation (BMIS) is an essential yet challenging task for the visual analysis of biomedical images. Modern deep learning-based architectures, such as UNet, UNet-based variants, Transformers-based networks, and their combinations, have achieved reasonable success in BMIS. However, they still face certain shortcomings in extracting fine-grained features. They are also limited by scenarios where the modeling of local and global feature representations needs to be optimized cor-rectly for spatial dependency in the decoding process, which can result in duplicate data utilization throughout the architecture. Besides, Transformer-based models lack inductive bias in addition to the complexity of the models. As a result, it can perform unsatisfactorily in a lesser biomedical image setting. This paper proposes a novel encode-decoder architecture named DOLG-NeXt, incorporating three major enhancements over the UNet-based variants. Firstly, we integrate squeeze and excitation network (SE -Net)-driven ConvNeXt stages as encoder backbone for effective feature extraction. Secondly, we employ a deep orthogonal fusion of local and global (DOLG) features module in the decoder to retrieve fine-grained contextual feature representations. Finally, we construct a SE-Net-like lightweight attention net-work alongside the DOLG module to provide refined target-relevant channel-based feature maps for decoding. To objectively validate the proposed DOLG-NeXt method, we perform extensive quantitative and qualitative analysis on four benchmark datasets from different biomedical image modalities: colono-scopy, electron microscopy, fluorescence, and retinal fundus imaging. DOLG-NeXt achieves a dice coeffi-cient score of 95.10% in CVC-ClinicDB, 95.80% in ISBI 2012, 94.77% in 2018 Data Science Bowl, and 84.88% in the DRIVE dataset, respectively. The experimental analysis shows that DOLG-NeXt outperforms several state-of-the-art models for BMIS tasks.(c) 2023 Elsevier B.V. All rights reserved.
引用
收藏
页数:14
相关论文
共 63 条
  • [31] DS-TransUNet: Dual Swin Transformer U-Net for Medical Image Segmentation
    Lin, Ailiang
    Chen, Bingzhi
    Xu, Jiayu
    Zhang, Zheng
    Lu, Guangming
    Zhang, David
    [J]. IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2022, 71
  • [32] Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
    Liu, Ze
    Lin, Yutong
    Cao, Yue
    Hu, Han
    Wei, Yixuan
    Zhang, Zheng
    Lin, Stephen
    Guo, Baining
    [J]. 2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021), 2021, : 9992 - 10002
  • [33] A ConvNet for the 2020s
    Liu, Zhuang
    Mao, Hanzi
    Wu, Chao-Yuan
    Feichtenhofer, Christoph
    Darrell, Trevor
    Xie, Saining
    [J]. 2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2022, : 11966 - 11976
  • [34] Lou AE, 2021, Arxiv, DOI arXiv:2105.04075
  • [35] MCRNet: Multi-level context refinement network for semantic segmentation in breast ultrasound imaging
    Louv, Meng
    Meng, Jie
    Qi, Yunliang
    Li, Xiaorong
    Ma, Yide
    [J]. NEUROCOMPUTING, 2022, 470 : 154 - 169
  • [36] Loss odyssey in medical image segmentation
    Ma, Jun
    Chen, Jianan
    Ng, Matthew
    Huang, Rui
    Li, Yu
    Li, Chen
    Yang, Xiaoping
    Martel, Anne L.
    [J]. MEDICAL IMAGE ANALYSIS, 2021, 71
  • [37] Oktay O, 2018, Arxiv, DOI arXiv:1804.03999
  • [38] Paul Bishmoy, 2020, COMPUT BIOL MED, V128
  • [39] Fine-Tuning CNN Image Retrieval with No Human Annotation
    Radenovic, Filip
    Tolias, Giorgos
    Chum, Ondrej
    [J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2019, 41 (07) : 1655 - 1668
  • [40] Ramachandran P, 2019, ADV NEUR IN, V32