基于三维循环残差卷积的脊柱CT图像分割

doi:10.19678/j.issn.1000-3428.0067751

摘要/Abstract

摘要： 脊柱计算机断层摄影(CT)图像的自动分割能够辅助医生诊疗相关疾病,相较于二维分割后再进行三维重建,三维分割方法更方便且能保留图像的空间信息。针对现有三维脊柱分割方法精度较低的问题,提出一种以三维循环残差卷积为基础的U型网络对脊柱CT图像进行分割。在网络前端引入三维坐标注意力机制使网络关注感兴趣的区域,使用三维循环残差模块代替普通卷积模块,使得网络在有效累积特征的同时缓解梯度消失问题。加入高效密集连接混合卷积模块减少底层细小特征信息的丢失,提出双特征残差注意力机制代替跳跃连接进行高低层级间的语义融合,通过聚合不同层级特征对全局上下文进行建模,提升分割性能。实验结果表明:在CSI2014公开数据集上,该网络Dice相似系数(DSC)达到93.85%,相较于对比的分割网络提升了1.77~7.65个百分点,相较于其他脊椎分割方法提升了1.67~10.85个百分点;在本地腰椎数据集上,相较于对比的分割模型DSC提升了1.51~19.86个百分点,验证了所提方法的有效性和应用于计算机辅助诊疗的可行性。

关键词: 脊柱分割, 三维医学图像, 深度学习, 注意力机制, 循环残差卷积

Abstract: The automatic segmentation of spine Computed Tomography(CT) images can assist doctors in diagnosing related diseases. Compared to Three-Dimensional(3D) reconstruction after Two-Dimensional(2D) segmentation, the 3D segmentation method is more convenient and can retain the spatial information of the image. To address the problem of the low accuracy of 3D spine segmentation, a U-Net based on 3D recurrent residual convolution to segment CT images of the spine is proposed in this study. A coordination attention mechanism is introduced in the network front to focus the network on the region of interest. A 3D recurrent residual module is used instead of a typical convolution module to accumulate features effectively and mitigate gradient disappearance. An efficient connected hybrid convolution module is added to preserve the tiny features. The dual-feature residual attention module is used instead of the jump connection for multiscale fusion to fuse semantics between high and low levels, and the global context is modeled by aggregating the features of different levels to improve the segmentation performance. First, the model is tested on the public datasets of CSI2014, and compared with other 3D segmentation networks and different spine segmentation methods, the Dice Similarity Coefficient(DSC) reaches 93.85%, which is 1.77-7.65 percentage points higher than those of other six segmentation networks and 1.67-10.85 percentage points higher than those of other spine segmentation methods. The model is also tested on the local lumbar dataset, and the DSC is increased by 1.51-19.86 percentage points compared with those of the other six segmentation models, verifying the effectiveness of the method proposed in this study and the feasibility of applying it to computer-aided diagnosis and treatment.

Key words: spine segmentation, Three-Dimensional(3D) medical image, deep learning, attention mechanism, recurrent residual convolution

中图分类号:

TP18

杨玉聃, 张俊华, 刘云凤. 基于三维循环残差卷积的脊柱CT图像分割[J]. 计算机工程, 2024, 50(4): 237-246.

YANG Yudan, ZHANG Junhua, LIU Yunfeng. Segmentation of Spine Computed Tomography Images Based on Three-Dimensional Recurrent Residual Convolution[J]. Computer Engineering, 2024, 50(4): 237-246.

https://www.ecice06.com/CN/Y2024/V50/I4/237

参考文献

[1] SHUVO M B, AHOMMED R, REZA S, et al. CNL-UNet:a novel lightweight deep learning architecture for multimodal biomedical image segmentation with false output suppression[J]. Biomedical Signal Processing and Control, 2021, 70:102959.
[2] RONNEBERGER O, FISCHER P, BROX T. U-Net:convolutional networks for biomedical image segmentation[EB/OL].[2023-04-02]. https://arxiv.org/abs/1505.04597.
[3] 于文涛, 张俊华, 梅建华, 等. 脊柱MR图像自动分割方法的研究[J]. 计算机工程与应用, 2022, 58(22):203-209. YU W T, ZHANG J H, MEI J H, et al. Research on automatic segmentation method of spinal MR images[J]. Computer Engineering and Applications, 2022, 58(22):203-209. (in Chinese)
[4] ZHOU Z W, SIDDIQUEE M M R, TAJBAKHSH N, et al. UNet++:redesigning skip connections to exploit multiscale features in image segmentation[J]. IEEE Transactions on Medical Imaging, 2020, 39(6):1856-1867.
[5] ZHANG L R, YANG J L, LIU D, et al. Spine X-ray image segmentation based on transformer and adaptive optimized postprocessing[C]//Proceedings of the 2nd International Conference on Software Engineering and Artificial Intelligence. Washington D. C., USA:IEEE Press, 2022:88-92.
[6] STRUDEL R, GARCIA R, LAPTEV I, et al. Segmenter:transformer for semantic segmentation[C]//Proceedings of IEEE/CVF International Conference on Computer Vision. Washington D. C., USA:IEEE Press, 2021:7262-7272.
[7] 周静, 钟原, 李平, 等. 用于颈椎MRI分割的多尺度特征融合注意力网络模型[J]. 计算机工程, 2023, 49(10):298-304. ZHOU J, ZHONG Y, LI P, et al. Multi-scale feature fusion attention network model for cervical vertebrae MRI segmentation[J]. Computer Engineering, 2023, 49(10):298-304. (in Chinese)
[8] ÇIÇEK Ö, ABDULKADIR A, LIENKAMP S S, et al. 3D U-Net:learning dense volumetric segmentation from sparse annotation[C]//Proceedings of the 19th International Conference on Medical Image Computing and Computer Assisted Intervention. Berlin, Germany:Springer, 2016:424-432.
[9] LIU Z, SU Z H, WANG M, et al. Computerized characterization of spinal structures on MRI and clinical significance of 3D reconstruction of lumbosacral intervertebral foramen[J]. Pain Physician, 2022, 25(1):27-35.
[10] LI W Q, TANG Y M, WANG Z Y, et al. Atrous residual interconnected encoder to attention decoder framework for vertebrae segmentation via 3D volumetric CT images[J]. Engineering Applications of Artificial Intelligence, 2022, 114:105102.
[11] TAO R, LIU W Y, ZHENG G Y. Spine-transformers:vertebra labeling and segmentation in arbitrary field-of-view spine CTs via 3D transformers[J]. Medical Image Analysis, 2022, 75:102258.
[12] LI T Y, WEI B Z, CONG J Y, et al. S³egANet:3D spinal structures segmentation via adversarial nets[J]. IEEE Access, 2020, 8:1892-1901.
[13] 刘侠, 甘权, 李冰, 等. 融合加权随机森林的自动3D椎骨CT图像主动轮廓分割方法[J]. 光电工程, 2020, 47(12):37-48. LIU X, GAN Q, LI B, et al. Automatic 3D vertebrae CT image active contour segmentation method based on weighted random forest[J]. Opto-Electronic Engineering, 2020, 47(12):37-48. (in Chinese)
[14] HU J, SHEN L, ALBANIE S, et al. Squeeze-and-excitation networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(8):2011-2023.
[15] WOO S, PARK J, LEE J Y, et al. CBAM:convolutional block attention module[C]//Proceedings of the 15th European Conference on Computer Vision. Berlin, Germany:Springer, 2018:3-19.
[16] HOU Q B, ZHOU D Q, FENG J S. Coordinate attention for efficient mobile network design[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA:IEEE Press, 2021:13713-13722.
[17] HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA:IEEE Press, 2016:770-778.
[18] LIANG M, HU X L. Recurrent convolutional neural network for object recognition[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA:IEEE Press, 2015:3367-3375.
[19] WANG P Q, CHEN P F, YUAN Y, et al. Understanding convolution for semantic segmentation[C]//Proceedings of IEEE Winter Conference on Applications of Computer Vision. Washington D. C., USA:IEEE Press, 2018:1451-1460.
[20] YAO J H, BURNS J E, MUNOZ H, et al. Detection of vertebral body fractures based on cortical shell unwrapping[C]//Proceedings of the 15th International Conference on Medical Image Computing and Computer Assisted Intervention. Berlin, Germany:Springer, 2012:509-516.
[21] RODRIGUEZ J D, PEREZ A, LOZANO J A. Sensitivity analysis of k-fold cross validation in prediction error estimation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32(3):569-575.
[22] MILLETARI F, NAVAB N, AHMADI S A. V-Net:fully convolutional neural networks for volumetric medical image segmentation[C]//Proceedings of the 4th International Conference on 3D Vision. Washington D. C., USA:IEEE Press, 2016:565-571.
[23] ISENSEE F, JAEGER P F, KOHL S A A, et al. nnU-Net:a self-configuring method for deep learning-based biomedical image segmentation[J]. Nature Methods, 2021, 18:203-211.
[24] CAI S J, TIAN Y X, LUI H, et al. Dense-UNet:a novel multiphoton in vivo cellular image segmentation model based on a convolutional neural network[J]. Quantitative Imaging in Medicine and Surgery, 2020, 10(6):1275-1285.
[25] OKTAY O, SCHLEMPER J, FOLGOC L L, et al. Attention U-Net:learning where to look for the pancreas[EB/OL].[2023-04-02].https://arxiv.org/pdf/1804.03999.pdf.
[26] HATAMIZADEH A, TANG Y C, NATH V, et al. UNETR:transformers for 3D medical image segmentation[C]//Proceedings of IEEE/CVF Winter Conference on Applications of Computer Vision. Washington D. C., USA:IEEE Press, 2022:574-584.
[27] SEITEL A, RASOULIAN A, ROHLING R, et al. Lumbar and thoracic spine segmentation using a statistical multi-object shape+pose model[M]//YAO J H, GLOCKER B, KLINDER T, et al. Recent advances in computational methods and clinical applications for spine imaging. Berlin, Germany:Springer, 2015:221-225.
[28] QADRI S F, ZHAO Z Q, AI D N, et al. Vertebrae segmentation via stacked sparse autoencoder from computed tomography images[C]//Proceedings of the 11th International Conference on Digital Image Processing. Washington D. C., USA:IEEE Press, 2019:1206-1211.
[29] LI B, LIU C, WU S Y, et al. Verte-Box:a novel convolutional neural network for fully automatic segmentation of vertebrae in CT image[J]. Tomography, 2022, 8(1):45-58.

选择文件类型/文献管理软件名称

选择包含的内容