Chunshi Wang

Chunshi Wang

Auto-generated profile, based on publicly available information

Chunshi Wang

School of Software Technology, Zhejiang University, Ningbo, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
18 September 2024 in IEEE Journal of Biomedical and Health Informatics
Verified
Journal Article
Closed access
BZBin Zhao
Bin Zhao
CWChunshi Wang
Chunshi Wang
Shuxue DingShuxue Ding
Shuxue Ding
Semi-supervised learning for medical image segmentation presents a unique challenge of efficiently using limited labeled data while leveraging abundant unlabeled data. Despite advancements, existing methods often do not fully exploit the potential of the unlabeled data for enhancing model robustness and accuracy. In this paper, we introduce CrossMatch, a novel framework that integrates knowledge distillation with dual perturbation strategies, image-level and feature-level, to improve the model's learning from both labeled and unlabeled data. CrossMatch employs multiple encoders and decoders to generate diverse data streams, which undergo self-knowledge distillation to enhance the consistency and reliability of predictions across varied perturbations. Our method significantly surpasses other state-of-the-art techniques in standard benchmarks by effectively minimizing the gap between training on labeled and unlabeled data and improving edge accuracy and generalization in medical image segmentation. The efficacy of CrossMatch is demonstrated through extensive experimental validations, showing remarkable performance improvements without increasing computational costs. Code for this implementation is made available at https://github.com/AiEson/CrossMatch.git .

5 July 2025 in Pattern Recognition
Verified
Journal Article
Open access
YWYaqi Wang
Yaqi Wang
YZYifan Zhang
Yifan Zhang
XCXiao-Diao Chen
Xiao-Diao Chen
SWShuai Wang
Shuai Wang
DQDahong Qian
Dahong Qian
Qun JinQun Jin
Qun Jin


25 April 2025 in Pattern Recognition Letters
Verified
Journal Article
Closed access
CWChunshi Wang
Chunshi Wang
CXChuan Xiong
Chuan Xiong
BZBin Zhao
Bin Zhao
Shuxue DingShuxue Ding
Shuxue Ding

20 October 2023
Verified
Conference Paper
Open access
BZBin Zhao
Bin Zhao
TYTing Yang
Ting Yang
ZXZhaowei Xin
Zhaowei Xin
CWChunshi Wang
Chunshi Wang
Computed tomography (CT) scanning is a common method for diagnosing rib fractures. However, the shape of the fracture region obtained from CT image segmentation seriously affects the clinician's diagnosis. To solve this problem, we propose a segmentation network (3D FM-UNet) to segment CT images to obtain more accurate regions of rib fractures. The experiments show that our 3D FM-UNet achieves an average mean Dice score of 0.5452 and an average IoU of 0.3748, which exceeds the comparison methods.

13 August 2024
Verified
Conference Paper
Open access
YWYifan Wang
Yifan Wang
HLHaidong Liang
Haidong Liang
CWChunshi Wang
Chunshi Wang
BZBin Zhao
Bin Zhao
How to maximise the unlabelled data to enhance the segmentation performance of the model is the major challenge for semi-supervised learning. Although current research has made some progress, existing techniques do not fully make use of the potential value of unlabelled data. To address this problem, this paper proposes a semi-supervised segmentation method that incorporates knowledge distillation and dual perturbation strategies to optimize the model's learning process for both labelled and unlabelled data. Extensive experiments on two benchmark datasets evaluate that our method outperforms the comparison methods.