MAA-Net：Gastric Tumor Segmentation and T Staging Algorithm

doi:10.3778/j.issn.1002-8331.2012-0489

Abstract

Abstract: Computed tomography（CT） technology plays an important role in the early screening, clinical diagnosis, preoperative prediction, postoperative evaluation of gastric diseases, and is an important basis for doctors to diagnose gastric diseases. Aiming at the problem of the large and complex structure of the stomach tissue, it is difficult to accurately segment and T stage the lesion, a multi-task convolutional neural network MAA-Net is proposed. This new method contains two main lines：One main line is used to segment gastric tumors in a multi-input U-shaped structure; the other main line uses dense cavity convolution modules to extract deep information for gastric cancer T staging. In addition, for the problem of low tumor segmentation accuracy, an adaptive feature fusion module is proposed. In order to improve the segmentation and gradient change of small targets, attention mechanism and hybrid loss function are respectively proposed. Quantitative and qualitative evaluation and analysis show that the proposed method is better than similar methods. If this method is used as a tool for early detection of gastric cancer, it can effectively relieve the pressure on doctors and help patients in time.

Key words: gastric tumor segmentation, T staging of gastric cancer, dense dilated convolution, adaptive feature fusion, fusion loss function, attention mechanism

摘要： 计算机断层扫描技术（computed tomography，CT）在胃部疾病的早期筛查、临床诊断、术前预测、术后评估等方面发挥重要作用，是医生诊断胃部疾病的重要依据。针对胃部组织形变大、结构复杂，难以精确地对病灶进行分割和T分期的问题，提出了一种多任务卷积神经网络MAA-Net。这种新型的方法包含两条主线：一条主线在多输入的U型结构中进行胃部肿瘤的分割；另一条主线采用密集空洞卷积模块提取深层的特征信息进行胃癌的T分期。针对肿瘤分割精度低的问题，提出了自适应特征融合模块。为了改善小目标的分割和梯度变化，分别提出了注意力机制和混合损失函数。对所提方法进行定量定性的评估和分析，结果表明，所提方法优于同类方法。这种方法若作为胃癌早期检测的工具，可以有效地缓解医生的压力并及时帮助患者。

关键词: 胃部肿瘤分割, 胃癌T分期, 密集空洞卷积, 自适应特征融合, 混合损失函数, 注意力机制

ZHOU Yilong, WEI Ziran, CAI Qingping, GAO Yongbin, MA Shuo. MAA-Net：Gastric Tumor Segmentation and T Staging Algorithm[J]. Computer Engineering and Applications, 2022, 58(16): 226-233.

周意龙, 卫子然, 蔡清萍, 高永彬, 马硕. MAA-Net：胃部肿瘤分割与T分期算法[J]. 计算机工程与应用, 2022, 58(16): 226-233.

References

[1] ETEMADI A，SAFIRI S，SEPANLOU S，et al.The global，regional，and national burden of stomach cancer in 195 countries 2017：a systematic analysis for the global burden of disease study 2017[J].The Lancet Gastroenterology Hepatology，2020，5（1）：42-54.
[2] GOTODA T.Endoscopic resection of early gastric cancer[J].Gastric Cancer，2007，10（1）：1-11.
[3] MAINTZ J B，VIERGEVER M A.A survey of medical image registration[J].Medical Image Analysis，1998，2（1）：1-36.
[4] LITJENS G J，KOOI T，BEJNORDI B E，et al.A survey on deep learning in medical image analysis[J].Medical Image Analysis，2017，42：60-88.
[5] MEHTA S，MERCAN E，BARTLETT J，et al.Y-Net：joint segmentation and classification for diagnosis of breast biopsy images[C]//Proceedings of the 21st International Conference on Medical Image Computing and Computer Assisted Intervention.Cham：Springer，2018：893-901.
[6] LECUN Y，BOTTOU L，BENGIO Y，et al.Gradient-based learning applied to document recognition[J].Proceedings of the IEEE，1998，86（11）：2278-2324.
[7] KRIZHEVSKY A，SUTSKEVER I，HINTON G E.ImageNet classification with deep convolutional neural networks[C]//Proceedings of the 25th International Conference on Neural Information Processing Systems.Red Hook：Curran Associates Inc，2012：1097-1105.
[8] SIMONYAN K，ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[J].arXiv：1409.
1556，2014.
[9] SZEGEDY C，LIU W，JIA Y Q，et al.Going deeper with convolutions[C]//Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition，2015：1-9.
[10] HE K，ZHANG X，REN S，et al.Deep residual learning for image recognition[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition，2016：770-778.
[11] HUANG G，LIU Z，VAN DER MAATEN L，et al.Densely connected convolutional networks[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition，Honolulu，2017：2261-2269.
[12] LONG J，SHELHAMER E，DARRELL T.Fully convolutional networks for semantic segmentation[C]//Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition，2015：3431-3440.
[13] RONNEBERGER O，FISCHER P，BROX T.U-Net：convolutional networks for biomedical image segmentation[C]//Proceedings of the 18th International Conference on International Conference on Medical Image Computing and Computer-Assisted Intervention.Cham：Springer，2015：234-241.
[14] GIBSON E，GIGANTI F，HU Y，et al.Automatic multi-organ segmentation on abdominal CT with dense V-networks[J].IEEE Transactions on Medical Imaging，2018，37（8）：1822-1834.
[15] FU H，CHENG J，XU Y，et al.Joint optic disc and cup segmentation based on multi-label deep network and polar transformation[J].IEEE Transactions on Medical Imaging，2018，37（7）：1597-1605.
[16] PENG C，ZHANG X，YU G，et al.Large kernel matters improve semantic segmentation by the global convolutional network[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition，2017：1743-1751.
[17] GU Z，CHENG J，FU H，et al.CE-Net：context encoder network for 2D medical image segmentation[J].IEEE Transactions on Medical Imaging，2019，38（10）：2281-2292.
[18] ZHAO H，SHI J，QI X，et al.Pyramid scene parsing network[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition，2017：2881-2890.
[19] SHIBATA T，TERAMOTO A，YAMADA H，et al.Automated detection and segmentation of early gastric cancer from endoscopic images using Mask R-CNN[J].Applied Sciences，2020，10（11）：3842.
[20] SUN M，ZHANG G，DANG H，et al.Accurate gastric cancer segmentation in digital pathology images using deformable convolution and multi-scale embedding networks[J].IEEE Access，2019，7：75530-75541.
[21] ZHENG L B，ZHANG X Y，HU J L，et al.Establishment and applicability of a diagnostic system for advanced gastric cancer T staging based on a faster region-based convolutional neural network[J].Frontiers in Oncology，2020，10：1238.
[22] LIN T Y，GOYAL P，GIRSHICK R，et al.Focal loss for dense object detection[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision，2017：2980-2988.
[23] SHORE J，JOHNSON R.Axiomatic derivation of the principle of maximum entropy and the principle of minimum cross-entropy[J].IEEE Transactions on Information Theory，1980，26（1）：26-37.
[24] MILLETARI F，NAVAB N，AHMADI S A，et al.V-Net：fully convolutional neural networks for volumetric medical image segmentation[C]//Proceedings of the 4th International Conference on 3D Vision，2016：565-571.
[25] KESKAR N S，SOCHER R.Improving generalization performance by switching from Adam to SGD[J].arXiv：1712.07628.2017.