基于多尺度注意力和数据增强的细胞核分割

doi:10.19678/j.issn.1000-3428.0069116

摘要/Abstract

摘要：

U-Net因结构简单且高效被广泛应用于医学分割领域。然而, U-Net的跳跃连接不能很好地弥补编码器和解码器之间的语义差距。而医学分割数据的标注要求严格, 使得数据集数量和规模都较小。针对上述问题, 设计多尺度注意力融合(MSAF)模块, 旨在利用注意力机制可调整网络学习方向的特点和多尺度特征融合来有效缓解语义偏差。MSAF模块在前2个阶段使用通道注意力来捕获全局特征; 在后2个阶段使用空间注意力来捕获局部特征; 最后将多个阶段提取的特征进行融合以增强特征信息。此外, 提出基于傅里叶变换的数据增强(FTDA)方法解决医学分割数据集稀少的问题。FTDA通过扰动输入图像在频域中的幅度信息实现其相位信息的数据增强。在MoNuSeg、CryoNuSeg和2018 Data Science Bowl数据集上的实验结果表明, 提出方法的mIoU和Dice指标比其他先进方法表现出更好的性能。此外, 提出的FTDA方法对小规模数据集也具有较好的增益效果。

关键词: 注意力机制, U-Net模型, 傅里叶变换, 细胞核分割, 数据增强

Abstract:

The U-Net has been widely used in medical segmentation because of its simplicity and efficiency. However, the skip connections in the U-Net do not adequately bridge the semantic gap between the encoder and decoder. In addition, the strict labeling requirements for medical segmentation data reduce the number and scale of available datasets. To address the aforementioned issues, a Multi-Scale Attention Fusion(MSAF) module is designed to effectively alleviate semantic bias by utilizing the ability of the attention mechanism to adjust the learning direction of the network and incorporate multi-scale feature fusion. In the first two stages of the U-Net, channel attention is employed in MSAF to capture global features. In the next two stages, spatial attention is used to capture local features. Finally, the features extracted from the multiple stages are fused to enhance feature information. Moreover, Fourier Transform Data Augmentation(FTDA), a data augmentation method based on Fourier transform, is introduced to overcome the scarcity of medical segmentation data. FTDA enhances phase data by disturbing the amplitude data of the input image in the frequency domain. Experimental results on the MoNuSeg, CryoNuSeg, and 2018 Data Science Bowl datasets show that the mean Intersection over Union (mIoU) and Dice metrics of the proposed method are better than those of other advanced methods. Furthermore, the proposed FTDA method displays remarkable performance gains, even on small-scale datasets.

Key words: attention mechanism, U-Net model, Fourier transform, nucleus segmentation, data augmentation

张兴鹏, 何东, 杨模, 叶杭滨. 基于多尺度注意力和数据增强的细胞核分割[J]. 计算机工程, 2025, 51(2): 387-396.

ZHANG Xingpeng, HE Dong, YANG Mo, YE Hangbin. Nucleus Segmentation Based on Multiscale Attention and Data Augmentation[J]. Computer Engineering, 2025, 51(2): 387-396.

https://www.ecice06.com/CN/Y2025/V51/I2/387

图/表 10

图1 U-Net改进架构

Fig.1 U-Net improved architecture

图2 MSAF结构

Fig.2 Structure of MSAF

图3 FTDA流程

Fig.3 FTDA procedure

图4 对比实验的可视化结果

Fig.4 Visualizations results of compare experimentals

图5 热力图的可视化分析

Fig.5 Visualization analysis of heat map

参考文献 31

1	RONNEBERGER O , FISCHER P , BROX T . U-Net: convolutional networks for biomedical image segmentation. Berlin, Germany: Springer, 2015.
2	ALOM M Z , YAKOPCIC C , HASAN M , et al. Recurrent residual U-Net for medical image segmentation. Journal of Medical Imaging, 2019, 6 (1): 014006. URL
3	XU Q , MA Z C , HE N , et al. DCSAU-Net: a deeper and more compact split-attention U-Net for medical image segmentation. Computers in Biology and Medicine, 2023, 154, 106626. doi: 10.1016/j.compbiomed.2023.106626
4	ISENSEE F , JAEGER P F , KOHL S A A , et al. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nature Methods, 2021, 18, 203- 211. doi: 10.1038/s41592-020-01008-z
5	LIN J W , LIAO X , YU L , et al. Res-UNet based optic disk segmentation in retinal image. Journal of Computers, 2020, 31 (3): 183- 194. URL
6	JHA D, RIEGLER M A, JOHANSEN D, et al. DoubleU-net: a deep convolutional neural network for medical image segmentation[C]//Proceedings of the 33rd IEEE International Symposium on Computer-Based Medical Systems. Washington D. C., USA: IEEE Press, 2020: 558-564.
7	LI Y X , WANG S , WANG J , et al. GT U-Net: a U-Net like group transformer network for tooth root segmentation. Berlin, Germany: Springer, 2021.
8	CAO H , WANG Y Y , CHEN J , et al. Swin-UNet: UNet-like pure transformer for Medical image segmentation. Berlin, Germany: Springer, 2023.
9	WANG H N , CAO P , WANG J Q , et al. UCTransNet: rethinking the skip connections in U-Net from a channel-wise perspective with transformer. Artificial Intelligence, 2022, 36 (3): 2441- 2449.
10	IBTEHAZ N , RAHMAN M S . MultiResUNet: rethinking the U-Net architecture for multimodal biomedical image segmentation. Neural Networks, 2020, 121, 74- 87. doi: 10.1016/j.neunet.2019.08.025
11	ZHOU Z W , SIDDIQUEE M M R , TAJBAKHSH N , et al. UNet++: redesigning skip connections to exploit multiscale features in image segmentation. IEEE Transactions on Medical Imaging, 2020, 39 (6): 1856- 1867. doi: 10.1109/TMI.2019.2959609
12	HUANG H M, LIN L F, TONG R F, et al. UNet 3+: a full-scale connected UNet for medical image segmentation[C]//Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing. Washington D. C., USA: IEEE Press, 2020: 1055-1059.
13	YANG Y C, SOATTO S. FDA: Fourier domain adaptation for semantic segmentation[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, USA: IEEE Press, 2020: 4084-4094.
14	COOLEY J W , TUKEY J W . An algorithm for the machine calculation of complex Fourier series. Mathematics of Computation, 1965, 19 (90): 297- 301. doi: 10.1090/S0025-5718-1965-0178586-1
15	LIN S, LIU N, NAZEMI M, et al. FFT-based deep learning deployment in embedded systems[C]// Proceedings of Design, Automation Test in Europe Conference Exhibition. Dresden, Germany: [s. n.], 2018: 1045-1050.
16	CHITSAZ K, HAJABDOLLAHI M, KARIMI N, et al. Acceleration of convolutional neural network using FFT-based split convolutions[EB/OL]. [2023-11-10]. https://arxiv.org/pdf/2003.12621.
17	ZHANG J, LIN Y, SONG Z, et al. Learning long term dependencies via Fourier recurrent units[C]// Proceedings of International Conference on Machine Learning. Stockholm, Sweden: [s. n.], 2018: 5810-5818.
18	LEE-THORP J, AINSLIE J, ECKSTEIN I, et al. FNet: mixing tokens with Fourier transforms[C]//Proceedings of Conference of the North American Chapter of the Association for Computational Linguistics. Stroudsburg, USA: Association for Computational Linguistics, 2022: 4296-4313.
19	RAO Y, ZHAO W, ZHU Z, et al. Global filter networks for image classification[C]//Proceedings of Advances in Neural Information Processing Systems. Cambridge, USA: MIT Press, 2021: 980-993.
20	RIAZ H U M, BENBARKA N, ZELL A. FourierNet: compact mask representation for instance segmentation using differentiable shape decoders[C]//Proceedings of the 25th International Conference on Pattern Recognition. Washington D. C., USA: IEEE Press, 2021: 7833-7840.
21	XIAO X, LIAN S, LUO Z M, et al. Weighted res-UNet for high-quality retina vessel segmentation[C]//Proceedings of the 9th International Conference on Information Technology in Medicine and Education. Hangzhou, China: [s. n.], 2018: 327-331.
22	GUAN S , KHAN A A , SIKDAR S , et al. Fully dense UNet for 2-D sparse photoacoustic tomography artifact removal. IEEE Journal of Biomedical and Health Informatics, 2020, 24 (2): 568- 576. doi: 10.1109/JBHI.2019.2912935
23	褚张晴晴, 钟志强, 颜子夜, 等. 基于特征融合与注意力机制的脑肿瘤分割算法. 计算机工程, 2023, 49 (10): 154- 161. doi: 10.19678/j.issn.1000-3428.0066311
	CHU Z , ZHONG Z Q , YAN Z Y , et al. Brain tumor segmentation algorithm based on feature fusion and attention mechanism. Computer Engineering, 2023, 49 (10): 154- 161. doi: 10.19678/j.issn.1000-3428.0066311
24	WOO S , PARK J , LEE J Y , et al. CBAM: convolutional block attention module. Berlin, Germany: Springer, 2018.
25	OKTAY O, SCHLEMPER J, FOLGOC L L, et al. Attention U-net: learning where to look for the pancreas[EB/OL].[2023-11-10]. https://arxiv.org/pdf/1804.03999.
26	CHEN J N, LU Y Y, YU Q H, et al. TransUNet: transformers make strong encoders for medical image segmentation[EB/OL].[2023-11-10]. https://arxiv.org/pdf/2102.04306.
27	LI Z H , LI Y X , LI Q D , et al. LViT: language meets vision transformer in medical image segmentation. IEEE Transactions on Medical Imaging, 2024, 43 (1): 96- 107. doi: 10.1109/TMI.2023.3291719
28	XU G P, ZHANG X, FANG Y, et al. LeVit-UNet: make faster encoders with transformer for biomedical image segmentation[EB/OL].[2023-11-10]. https://arxiv.org/pdf/2107.08623.
29	CAICEDO J C , GOODMAN A , KARHOHS K W , et al. Nucleus segmentation across imaging experiments: the 2018 Data Science Bowl. Nature Methods, 2019, 16, 1247- 1253. doi: 10.1038/s41592-019-0612-7
30	KUMAR N , VERMA R , ANAND D , et al. A multi-organ nucleus segmentation challenge. IEEE Transactions on Medical Imaging, 2020, 39 (5): 1380- 1391. doi: 10.1109/TMI.2019.2947628
31	MAHBOD A , SCHAEFER G , BANCHER B , et al. CryoNuSeg: a dataset for nuclei instance segmentation of cryosectioned H & E-stained histological images. Computers in Biology and Medicine, 2021, 132, 104349. doi: 10.1016/j.compbiomed.2021.104349

[1]	张欢, 王晨, 单景东, 仇润鹤. 基于领域自适应与注意力机制的电梯安全风险预测[J]. 计算机工程, 2025, 51(2): 86-93.
[2]	张元, 吕德芳, 孟建军, 祁文哲. 基于双注意力和GSSN轻量化的钢轨扣件缺陷检测[J]. 计算机工程, 2025, 51(2): 289-299.
[3]	许明, 屈泰澎, 姜彦吉. 改进YOLOv7在复杂场景下的交通标志检测算法[J]. 计算机工程, 2025, 51(2): 335-343.
[4]	罗旭东, 袁笛, 常晓军, 何震宇. 基于不确定性启发图像增强的水下目标跟踪[J]. 计算机工程, 2025, 51(1): 11-19.
[5]	周雪阳, 傅启明, 陈建平, 陈延明, 陆悠, 王蕴哲. 基于证据和图推理的文档级关系抽取方法: 以医学关系为例[J]. 计算机工程, 2025, 51(1): 106-117.
[6]	肖超恩, 李子凡, 张磊, 王建新, 钱思源. 基于Transformer模型与注意力机制的差分密码分析[J]. 计算机工程, 2025, 51(1): 156-163.
[7]	杨红菊, 吉昌. 学习驱动的图像压缩算法研究[J]. 计算机工程, 2025, 51(1): 190-197.
[8]	胡涌涛, 黄洪琼. 结合特征融合和通道注意力的多分支换装行人重识别[J]. 计算机工程, 2025, 51(1): 225-234.
[9]	火久元, 苏泓瑞, 武泽宇, 王婷娟. 基于改进YOLOv8的道路交通小目标车辆检测算法[J]. 计算机工程, 2025, 51(1): 246-257.
[10]	郑雅洲, 刘万平, 黄东. 一种基于注意力机制的BERT-CNN-GRU检测方法[J]. 计算机工程, 2025, 51(1): 258-268.
[11]	王骞, 张俊华, 王泽彤, 李博. X2S-Net: 基于双平面X线片的脊柱三维重建[J]. 计算机工程, 2025, 51(1): 277-286.
[12]	刘钟, 唐宏, 王宁喆, 朱传润. 融合RNN与稀疏自注意力的文本摘要方法[J]. 计算机工程, 2025, 51(1): 312-320.
[13]	刘兆伟, 方艳红, 郑明宇, 锁斌. 基于注意力机制与多任务的肺部疾病诊断方法[J]. 计算机工程, 2025, 51(1): 332-342.
[14]	李维刚, 厉许昌, 田志强, 李金灵. 基于自蒸馏框架的点云分类及其鲁棒性研究[J]. 计算机工程, 2024, 50(9): 72-81.
[15]	李俊俊, 董建刚, 李坤. 基于Kubernetes的集群节能策略研究[J]. 计算机工程, 2024, 50(9): 82-91.

选择文件类型/文献管理软件名称

选择包含的内容