CTH-Net: A CNN and Transformer hybrid network for skin lesion segmentation

Yuhan Ding; Zhenglin Yi; Jiatong Xiao; Minghui Hu; Yu Guo; Zhifang Liao; Yongjie Wang

doi:10.1016/j.isci.2024.109442

CTH-Net: A CNN and Transformer hybrid network for skin lesion segmentation

iScience. 2024 Mar 6;27(4):109442. doi: 10.1016/j.isci.2024.109442. eCollection 2024 Apr 19.

Authors

Yuhan Ding¹, Zhenglin Yi², Jiatong Xiao², Minghui Hu², Yu Guo^{3

4}, Zhifang Liao¹, Yongjie Wang^{3

4}

Affiliations

¹ School of Computer Science and Engineering, Central South University, Changsha 410083, China.
² Departments of Urology, Xiangya Hospital, Central South University, Changsha 410008, China.
³ Department of Burns and Plastic Surgery, Xiangya Hospital, Central South University, Changsha 410008, China.
⁴ National Clinical Research Center for Geriatric Disorders, Xiangya Hospital, Central South University, Changsha 410008, China.

Abstract

Automatically and accurately segmenting skin lesions can be challenging, due to factors such as low contrast and fuzzy boundaries. This paper proposes a hybrid encoder-decoder model (CTH-Net) based on convolutional neural network (CNN) and Transformer, capitalizing on the advantages of these approaches. We propose three modules for skin lesion segmentation and seamlessly connect them with carefully designed model architecture. Better segmentation performance is achieved by introducing SoftPool in the CNN branch and sandglass block in the bottleneck layer. Extensive experiments were conducted on four publicly accessible skin lesion datasets, ISIC 2016, ISIC 2017, ISIC 2018, and PH² to confirm the efficacy and benefits of the proposed strategy. Experimental results show that the proposed CTH-Net provides better skin lesion segmentation performance in both quantitative and qualitative testing when compared with state-of-the-art approaches. We believe the CTH-Net design is inspiring and can be extended to other applications/frameworks.

Keywords: Artificial intelligence; Health sciences; Machine learning.