| 324 | 0 | 99 |
| 下载次数 | 被引频次 | 阅读次数 |
针对小样本医学图像分割的现有方法忽略了支持集和查询集之间的分布偏移和局部边缘细节的问题，提出了一种用于小样本医学图像分割的原型优化和细化分割网络(PORSNet)。网络中包含一个原型循环迭代模块，通过迭代执行初始原型校正、原型全局感知、原型蒸馏等步骤，抑制初始原型和查询集之间的分布偏移及增强原型的表达能力。此外，还包含一个原型细化分割模块，通过掩码引导聚合和特征归一细化，进一步处理边缘细节信息。在两个广泛使用的小样本医学公共数据集ABD-MRI和ABD-CT上进行大量验证表明，笔者提出的PORSNet可以利用少量样本在确保分割效果的同时具有良好的泛化能力。
Abstract：In order to solve the problem of few-shot medical image segmentation with the distribution shift and local edge details between the support set and the query set, a prototype optimization and refinement segmentation network(PORSNet) for few-shot medical image segmentation was proposed. The network contained a prototype loop iteration module, which suppressed the distribution shift between the initial prototype and the query set and enhanced the expressiveness of the prototype by iteratively performing steps such as initial prototype correction, prototype global perception, and prototype distillation. In addition, a prototype refinement segmentation module was included to further process edge detail information through mask-guided aggregation and feature normalization refinement. Extensive experiments on two widely used few-shot medical public datasets, ABD-MRI and ABD-CT, showed that the proposed PORSNet could use a small number of samples to ensure the segmentation effect with good generalization ability.
[1] HAN K,SHENG V S,SONG Y Q,et al.Deep semi-supervised learning for medical image segmentation:a review[J].Expert systems with applications,2024,245:123052.
[2] 郭婧，王飞.多尺度特征融合与交叉指导的小样本语义分割[J].中国图象图形学报，2024,29(5):1265-1276.GUO J,WANG F.Multiscale feature fusion and cross-guidance for few-shot semantic segmentation[J].Journal of image and graphics,2024,29(5):1265-1276.
[3] GUHA ROY A,SIDDIQUI S,PÖLSTERL S,et al.′Squeeze & excite′ guided few-shot segmentation of volumetric images[J].Medical image analysis,2020,59:101587.
[4] WANG K X,LIEW J H,ZOU Y T,et al.PANet:few-shot image semantic segmentation with prototype alignment[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV).Piscataway:IEEE Press,2019:9196-9205.
[5] OUYANG C,BIFFI C,CHEN C,et al.Self-supervision with superpixels:training few-shot medical image segmentation without annotation[M]//Lecture Notes in Computer Science.Cham:Springer International Publishing,2020:762-780.
[6] HANSEN S,GAUTAM S,JENSSEN R,et al.Anomaly detection-inspired few-shot medical image segmentation through self-supervision with supervoxels[J].Medical image analysis,2022,78:102385.
[7] WU H,XIAO F,LIANG C.Dual contrastive learning with anatomical auxiliary supervision for few-shot medical image segmentation[C]//European Conference on Computer Vision.Cham:Springer Press,2022:417-434.
[8] 贾熹滨，郭雄，王珞，等.一种迭代边界优化的医学图像小样本分割网络[J].自动化学报，2024,50(10):1988-2001.JIA X B,GUO X,WANG L,et al.A few-shot medical image segmentation network with iterative boundary refinement[J].Acta automatica sinica,2024,50(10):1988-2001.
[9] DING H,SUN C C,TANG H,et al.Few-shot medical image segmentation with cycle-resemblance attention[C]//2023 IEEE/CVF Winter Conference on Applications of Computer Vision.Piscataway:IEEE Press,2023:2487-2496.
[10] SHEN Q Q,LI Y N,JIN J Y,et al.Q-net:query-informed few-shot medical image segmentation[M]//Lecture Notes in Networks and Systems.Cham:Springer Press,2024:610-628.
[11] 魏明军，魏帅，刘亚志，等.基于跨层级注意力学习的RGB-T显著目标检测[J].郑州大学学报(理学版),2025,57(3):42-48.WEI M J,WEI S,LIU Y Z,et al.RGB-T salient object detection based on cross-level attention learning[J].Journal of Zhengzhou university(natural science edition),2025,57(3):42-48.
[12] LIN S H,XIE H W,WANG B,et al.Knowledge distillation via the target-aware transformer[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE Press,2022:10905-10914.
[13] YU T,GUO Z Y,JIN X,et al.Region normalization for image inpainting[J].Proceedings of the AAAI conference on artificial intelligence,2020,34(7):12733-12740.
基本信息:
DOI：10.13705/j.issn.1671-6841.2024125
中图分类号:R318;TP391.41
引用信息:
[1]魏明军,贺海鹏,陈伟彬,等.一种原型优化和细化分割的小样本医学图像分割网络[J].郑州大学学报(理学版),2026,58(02):10-16.DOI:10.13705/j.issn.1671-6841.2024125.
基金信息:
河北省高等学校科学技术研究项目(ZD2022102)
2025-02-08
2025-02-08
2025-02-08