Automatic segmentation of choroid by TransGLnet integrating attention mechanism

HUANG Wenbo; QU Chaofan; YAN Yang

doi:10.37188/OPE.20233123.3482

您当前的位置：

首页 >

文章列表页 >

Automatic segmentation of choroid by TransGLnet integrating attention mechanism

Information Sciences | 更新时间：2023-12-13

- Automatic segmentation of choroid by TransGLnet integrating attention mechanism
- Optics and Precision Engineering Vol. 31, Issue 23, Pages: 3482-3489(2023)
- 作者机构：
  
  长春师范大学计算机科学与技术学院，吉林长春 130032
- 作者简介：
- 基金信息：
- DOI：10.37188/OPE.20233123.3482
  CLC： TP391.7
- Received：15 June 2023，
  
  Revised：24 July 2023，
  
  Published：10 December 2023
- 稿件说明：
移动端阅览
黄文博,屈超凡,燕杨.融合注意力机制的TransGLnet脉络膜自动分割[J].光学精密工程,2023,31(23):3482-3489.

HUANG Wenbo,QU Chaofan,YAN Yang.Automatic segmentation of choroid by TransGLnet integrating attention mechanism[J].Optics and Precision Engineering,2023,31(23):3482-3489.
黄文博,屈超凡,燕杨.融合注意力机制的TransGLnet脉络膜自动分割[J].光学精密工程,2023,31(23):3482-3489. DOI： 10.37188/OPE.20233123.3482.

HUANG Wenbo,QU Chaofan,YAN Yang.Automatic segmentation of choroid by TransGLnet integrating attention mechanism[J].Optics and Precision Engineering,2023,31(23):3482-3489. DOI： 10.37188/OPE.20233123.3482.

摘要

针对脉络膜与巩膜间对比度低，脉络膜分割存在脉络膜下边界模糊，难以界定等问题，提出了融合注意力机制的TransGLnet脉络膜自动分割网络，在卷积层引入全局注意力模块（Global Attention Module，GAM），在特征之间应用矩阵乘法，在整体空间位置的多个特征之间建立非线性交互，在不使用大量参数的情况下提取全局特征；在卷积层和Transformer编码器之间引入局部注意力模块（Local Attention Module，LAM），以1/4特征图为基本单元探索局部特征，特征图元素位置移动规则为保持行位置的元素不变，将列位置的元素由大到小重新排列。两模块融合可令网络有效兼顾全局与局部特征。实验结果表明，TransGLnet网络的Dice值为0.91，准确率为0.98，平均交并比为0.89，F1值为0.90，豪斯多夫距离为6.56。与现有脉络膜自动分割方法相比，本文方法的各项性能指标均有提高。TransGLnet脉络自动分割网络具有较好的稳定性，可供临床借鉴。

Abstract

Addressing the challenge posed by the low contrast between the choroid and sclera in choroid segmentation， this research introduces the TransGLnet choroid automatic segmentation network， employing an attention mechanism. The incorporation of a Global Attention Module （GAM） within the convolutional layer involves matrix multiplication between features， establishing nonlinear interactions across multiple features in the global spatial context. This enables the extraction of global features without an excessive number of parameters. To explore local features， a Local Attention Module （LAM） is introduced between the convolution layer and Transformer encoder， focusing on a 1/4 feature graph. The movement rule for feature graph elements maintains row position consistency while rearranging elements in the column position from largest to smallest. The integration of these two modules ensures that the network effectively considers both global and local features. Experimental results showcase the efficacy of the proposed TransGLnet network with a Dice value of 0.91， accuracy at 0.98， equal crossover ratio of 0.89， F1 value reaching 0.90， and a Hausdorff distance of 6.56. Comparative analysis against existing automatic choroidal segmentation methods reveals notable improvements in performance metrics. The network presented in this study demonstrates robustness and stability， rendering it suitable for clinical reference.

关键词

Keywords

references

NICKLA D L ， WALLMAN J . The multifunctional choroid ［J］. Progress in Retinal and Eye Research ， 2010 ， 29 （ 2 ）： 144 - 168 . doi: 10.1016/j.preteyeres.2009.12.002 http://dx.doi.org/10.1016/j.preteyeres.2009.12.002

ALONSO-CANEIRO D ， READ S A ， COLLINS M J . Automatic segmentation of choroidal thickness in optical coherence tomography ［J］. Biomedical Optics Express ， 2013 ， 4 （ 12 ）： 2795 - 2812 . doi: 10.1364/boe.4.002795 http://dx.doi.org/10.1364/boe.4.002795

TIAN J ， MARZILIANO P ， BASKARAN M ， et al . Automatic measurements of choroidal thickness in EDI-OCT images ［C］. 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society. August 28-September 1 ， 2012 . San Diego， CA. IEEE ， 2012： 5360 - 5363 . doi: 10.1109/embc.2012.6347205 http://dx.doi.org/10.1109/embc.2012.6347205

YAN Q F ， GU Y Y ， ZHAO J Y ， et al . Automatic choroid layer segmentation in OCT images via context efficient adaptive network ［J］. Applied Intelligence ， 2023 ， 53 （ 5 ）： 5554 - 5566 .

许祥丛，陈俊彦，王雪花，等 . 通过SEC-UNet精准分割糖尿病视网膜病变眼底OCT图像脉络膜层［J］. 生物化学与生物物理进展， 2022 ， 49 （ 12 ）： 2450 - 2457 .

XU X C ， CHEN J Y ， WANG X H ， et al . Accurate segmentation of choroidal layer in OCT images of diabetic retinopathy by SEC-UNet ［J］. Progress in Biochemistry and Biophysics ， 2022 ， 49 （ 12 ）： 2450 - 2457 . （in Chinese）

ZHU L ， LI J M ， ZHU R L ， et al . Synergistically segmenting choroidal layer and vessel using deep learning for choroid structure analysis ［J］. Physics in Medicine and Biology ， 2022 ， 67 （ 8 ）. DOI： 10.1088/1361-6560/ac5ed7 http://dx.doi.org/10.1088/1361-6560/ac5ed7 .

KHAING T T ， OKAMOTO T ， YE C ， et al . ChoroidNET： a dense dilated U-net model for choroid layer and vessel segmentation in optical coherence tomography images ［J］. IEEE Access ， 2021 ， 9 ： 150951 - 150965 . doi: 10.1109/access.2021.3124993 http://dx.doi.org/10.1109/access.2021.3124993

XU X C ， WANG X H ， LIN J Y ， et al . Automatic segmentation and measurement of choroid layer in high myopia for OCT imaging using deep learning ［J］. Journal of Digital Imaging ， 2022 ， 35 （ 5 ）： 1153 - 1163 . doi: 10.1007/s10278-021-00571-x http://dx.doi.org/10.1007/s10278-021-00571-x

WU W J ， GONG Y ， HAO H Y ， et al . Choroidal layer segmentation in OCT images by a boundary enhancement network ［J］. Frontiers in Cell and Developmental Biology ， 2022 ， 10 ： 1060241 . doi: 10.3389/fcell.2022.1060241 http://dx.doi.org/10.3389/fcell.2022.1060241

WANG X L ， GIRSHICK R ， GUPTA A ， et al . Non-local neural networks ［C］. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. June 18 - 23 ， 2018 . Salt Lake City， UT， USA. IEEE ， 2018 ： 7794 - 7803 .

XU L ， WANG L ， LI Y ， et al . Big Model and Small Model： remote modeling and local information extraction module for medical image segmentation ［J］. Applied Soft Computing ， 2023 ， 136 ： 110128 . doi: 10.1016/j.asoc.2023.110128 http://dx.doi.org/10.1016/j.asoc.2023.110128

CHEN J ， LU Y ， YU Q ， et al . TransUNet ： Transformers Make Strong Encoders for Medical Image Segmentation ［EB/OL］. 2021 ： arXiv ： 2102 . 04306 . https：//arxiv.org/abs/2102.04306.pdf https://arxiv.org/abs/2102.04306.pdf .

HE K M ， ZHANG X Y ， REN S Q ， et al . Deep residual learning for image recognition ［C］. 2016 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）. June 27 - 30 ， 2016 . Las Vegas， NV， USA. IEEE ， 2016 ： 770 - 778 .

MILLETARI F ， NAVAB N ， AHMADI S A . V - net ： Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation ［EB/OL］. 2016 ： arXiv ： 1606 . 04797 . https：//arxiv.org/abs/1606.04797.pdf https://arxiv.org/abs/1606.04797.pdf . doi: 10.1109/3dv.2016.79 http://dx.doi.org/10.1109/3dv.2016.79

KERMANY D ， ZHANG K ， GOLDBAUM M . Labeled optical coherence tomography （oct） and chest x-ray images for classification ［J］. Mendeley data ， 2018 ， 2 （ 2 ）： 651 .

RONNEBERGER O , FISCHER P , BROX T . U-net: Convolutional networks for biomedical image segmentation [C]. Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015 : 18th International Conference, Munich, Germany , October 5 - 9 , 2015, Proceedings,Part III 18. Springer International Publishing, 2015: 234 - 241 . doi: 10.1007/978-3-319-24574-4_28 http://dx.doi.org/10.1007/978-3-319-24574-4_28

WANG S , LI L , ZHUANG X . AttU-Net: attention U-Net for brain tumor segmentation [C]. International MICCAI Brainlesion Workshop. Cham: Springer International Publishing , 2021 : 302 - 311 . doi: 10.1007/978-3-031-09002-8_27 http://dx.doi.org/10.1007/978-3-031-09002-8_27

ALOM M Z , HASAN M , YAKOPCIC C , et al . Recurrent residual convolutional neural network based on u-net (r2u-net) for medical image segmentation [J]. arXiv preprint arXiv : 1802.06955 , 2018 . doi: 10.1109/naecon.2018.8556686 http://dx.doi.org/10.1109/naecon.2018.8556686

ZUO Q , CHEN S , WANG Z . R2AU-Net: attention recurrent residual convolutional neural network for multimodal medical image segmentation [J]. Security and Communication Networks , 2021 , 2021 : 1 - 10 . doi: 10.1155/2021/6625688 http://dx.doi.org/10.1155/2021/6625688

CHEN J , LU Y , YU Q , et al . Transunet: Transformers make strong encoders for medical image segmentation [J]. arXiv preprint arXiv： 2102.04306 , 2021 .

CAO H , WANG Y , CHEN J , et al . Swin-unet: Unet-likeic pure transformer for medical image segmentation [C]. European Conference on Computer Vision. Cham ： Springer Nature Switzerland , 2022 : 205 - 218 . doi: 10.1007/978-3-031-25066-8_9 http://dx.doi.org/10.1007/978-3-031-25066-8_9

JAEGER S ， CANDEMIR S ， ANTANI S ， et al . Two public chest X-ray datasets for computer-aided screening of pulmonary diseases ［J］. Quantitative Imaging in Medicine and Surgery ， 2014 ， 4 （ 6 ）： 475 .

Views

496

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

3DRes-ViT knee osteoarthritis classification model based on multimodal fusion

Digital diffraction method for single retinal image enhancement

Segmentation network for metastatic lymph nodes of head and neck tumors

Pneumonia aided diagnosis model based on dense dual-stream focused network

Improved BYOL method for predicting epidermal growth factor receptor gene mutations in non-small cell lung cancer

Related Author

WANG Xin

CAI Xiaodong

XU Rui

SONG Yu

CAO Liangcai

ZHANG Shuhe

ZHOU Tao

ZHOU Zhongwei

Related Institution

College of Computer Science and Engineering， Changchun University of Technology

Information Department， Jilin Qianwei Hospital

Department of Precision Instruments， Tsinghua University

College of Computer Science and Engineering， North Minzu University

Key Laboratory of Image and Graphics Intelligent Processing of State Ethnic Affairs Commission， North Minzu University

AI问答

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China Postal code：130033
Tel：0431-86176855 Email：gxjmgc@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰