D-TransUNet: A Breast Tumor Ultrasound Image Segmentation Model Based on Deep Feature Fusion

Research Articie March 18,2024

Yiying Wan ¹ , Yaru Yang ² , Hongjiang Guo ³ , Yangtian Yan ⁴ , Tongtong Liu ⁵ , Wenpei Liu ⁶ , Yiru Wang ⁷ , Wenhang Wang ⁸ , Hao Dang ⁹ hide author's information

1 : School of Information Technology, Henan University of Chinese Medicine, Zhengzhou 450046
2 : School of Information Technology, Henan University of Chinese Medicine, Zhengzhou 450046
3 : School of Information Technology, Henan University of Chinese Medicine, Zhengzhou 450046
4 : School of Information Technology, Henan University of Chinese Medicine, Zhengzhou 450046
5 : School of Information Technology, Henan University of Chinese Medicine, Zhengzhou 450046
6 : School of Information Technology, Henan University of Chinese Medicine, Zhengzhou 450046
7 : School of Information Technology, Henan University of Chinese Medicine, Zhengzhou 450046
8 : School of Information Technology, Henan University of Chinese Medicine, Zhengzhou 450046
9 : School of Information Technology, Henan University of Chinese Medicine; Zhengzhou Key Laboratory of Intelligent Analysis and Utilization of Traditional Chinese Medicine Information, Zhengzhou

DOI: 10.55578/joaims.240226.001

Keywords: Breast tumor ultrasound image segmentation; Central dense connection; TransUNet; Deep learning

Cite this article: Wan, Yiying et al., D-TransUNet: A Breast Tumor Ultrasound Image Segmentation Model Based on Deep Feature Fusion, Journal of Artificial Intelligence for Medical Sciences, 5(1-2), 1-8, 2024.

Full Text

PDF

Download Citation

Abstract

Breast ultrasound is a widely utilized modality for breast cancer screening since its noninvasive, radiation-free, low-cost, and easy-to-operate characteristics. The segmentation of breast tumor ultrasound images aims to accurately delineate the lesion area, thereby enhancing the usability and reliability of auxiliary diagnosis. In the realm of deep learning, U-Net and its variants based on fully convolutional networks have demonstrated outstanding performance in various medical image segmentation tasks.