Jintao Duan¹, Miao Zhou², Yuxiang Wang¹, Fangfang Chen¹, Liangqing Lin³, Qinghua Wu³, Haipo Cui¹

¹School of Health Science and Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China. ²Jiangsu Cancer Hospital, Nanjing 213164, Jiangsu Province, China. ³The First Hospital of Putian, Putian 351100, Fujian Province, China.

Address correspondence to: Haipo Cui, School of Health Science and Engineering, University of Shanghai for Science and Technology, No. 516 Jungong Road, Yangpu District, Shanghai 200093, China. E-mail: h_b_cui@163.com.

Received September 27, 2024; Accepted January 22, 2025; Published April 1, 2025

DOI: https://doi.org/10.61189/242224ubhmhy

Highlights

● A systematic review of deep learning (DL) approaches for femoral cartilage segmentation in knee joint ultrasound images.

● Evaluation of popular DL models (U-Net, U-Net++, Siam U-Net, Mask R-CNNs) using various metrics.

● Discussion on dataset scarcity, data preprocessing methods, and future directions for DL-based ultrasound segmentation.

Fangfang Chen¹, Miao Zhou², Jintao Duan¹, Yuxiang Wang¹, Liangqing Lin⁴, Wenhui Guo³, Yongchu Hu⁵

¹School of Health Science and Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China. ²Jiangsu Cancer Hospital, Nanjing 213164, Jiangsu Province, China. ³School of Anesthesiology, Naval Medical University, Shanghai 200433, China. ⁴The First Hospital of Putian, Putian 351100, Fujian Province, China. ⁵The Department of Anesthesiology, Long March Hospital, Shanghai 200003, China.

Address correspondence to: Wenhui Guo, School of Anesthesiology, Naval Medical University, No.25 Zhongyuan Road, Shanghai 200093, China. E-mail: wendyguo17@outlook.com. Yongchu Hu, The De-partment of Anesthesiology, Long March Hospital, No.415 Fengyang Road, Shanghai 200003, China. E-mail address: Adsfoxcn@sina.com.cn.

DOI: https://doi.org/10.61189/341921wbvxvz

Received September 27, 2024; Accepted January 22, 2025; Published April 1, 2025

Highlights

● Comparison of Methods: This article compares traditional techniques, such as thresholding and edge detection, with deep learning-based methods for breast ultrasound segmentation.

● U-Net’s Effectiveness: U-Net is identified as the benchmark for medical image segmentation due to its efficiency and ability to preserve details.

● Advantages of Deep Learning: Deep learning models, like CNNs and FCNs, improve segmentation accuracy by learning directly from raw data, while also mitigating noise.

Yuxiang Wang¹, Miao Zhou², Fangfang Chen¹, Jintao Duan¹, Liangqing Lin³, Qinghua Wu³, Wenhui Guo⁴, Haipo Cui¹

¹School of Health Science and Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China. ²Jiangsu Cancer Hospital, Nanjing 213164, China. ³Anesthesiology, The First Hospital of Putian, Putian 351100, China. ⁴School of Anesthesiology, Second Military Medical University/Naval Medical University, Shanghai 200433, China.

Address correspondence to: Haipo Cui, School of Health Science and Engineering, University of Shanghai for Science and Technology, No. 516 Jungong Road, Yangpu District, Shanghai 200093, China. E-mail: h_b_cui@163.com.

Received September 27, 2024; Accepted January 22, 2025; Published April 1, 2025

DOI: https://doi.org/10.61189/861515qdddmg

Highlights

● This review provides a comprehensive overview of various U-Net architectures and their variants, including the original U-Net, U-Net++, Attention U-Net, and ResU-Net, along with a discussion on potential improvements to the

U- Net architecture.

● The differences between these network architectures are analyzed in terms of training complexity and computa-   tional requirements.

● The review delves into the application of U-Net and its variants in ultrasound imaging, discussing both the advan-   tages and limitations of each model in various ultrasound contexts. Relevant literature on the application of each network architecture in ultrasound is also summarized.