Study of Ultrasonic Thyroid Nodules Detection Based on Cascade Rcnn
Zhang Haowei1*, Li Zhanqi1, Liu Ying1, Li Miao2
1(School of Health Science and Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China) 2(Zhengzhou Yihe Hospital of Henan Province, Zhengzhou 450000, China)
Abstract：Thyroid ultrasound images have low contrast, unclear edges, high noise, and the surrounding tissues are complex and difficult to distinguish, making it extremely difficult for doctors to diagnose thyroid diseases. To overcome this problem, Cascade Rcnn target detection algorithm was used in this work, with ResNet50, Resnet101 and fusion compression incentive attention modules SE-ResNet50, SE-ReNet101 as the backbone network. There were 1 513 cases thyroid ultrasound images (including 832 cases benign nodules and 681 cases malignant nodules) obtained from a third-class hospital. Under the guidance of professional sonographers, the data were preprocessed into the standard coco format data set. The weights obtained from the pre-training of the large Imagenet database by transfer learning were migrated to this experimental model structure. Comparing with the experimental results of the four backbone networks, Cascade Rcnn algorithm with SE-ResNet101 as the backbone network achieved an accuracy of 92.4%, recall rate of 86.2%, specificity of 95.1%1, F1 value of 89.22%, and mAP value of 82.4%. The detection result of nodule localization and classification of benign and malignant was of clinical guiding significance for assisting doctors in the diagnosis of thyroid ultrasound images.
章浩伟, 李占齐, 刘颖, 李淼. 基于Cascade Rcnn的超声甲状腺结节检测研究[J]. 中国生物医学工程学报, 2022, 41(1): 64-72.
Zhang Haowei, Li Zhanqi, Liu Ying, Li Miao. Study of Ultrasonic Thyroid Nodules Detection Based on Cascade Rcnn. Chinese Journal of Biomedical Engineering, 2022, 41(1): 64-72.
[1] 王昊.甲状腺结节超声图像纹理特征提取及半监督分级方法研究[D]. 成都: 西南交通大学, 2017. [2] 周姝君. 基于深度学习的超声甲状腺结节诊断研究[D]. 成都: 电子科技大学, 2019. [3] Xiao Ting, Liu Lei, Li Kai, et al. Comparison of Transferred Deep Neural Networks in Ultrasonic Breast Masses Discrimination[J]. BioMed Research International, 2018, 2018:1-9. [4] 吴迪. 甲状腺结节诊断分类方法的研究与实现[D]. 上海: 东华大学, 2017. [5] 熊伟, 龚勋, 罗俊,等. 基于局部纹理特征的超声甲状腺结节良恶性识别[J]. 数据采集与处理, 2015, 30(1):186-191. [6] 张振宇. 改进集成技术在甲状腺超声图像分类中的应用研究[D]. 保定: 河北大学, 2014. [7] Mohammadi SM, Helfroush MS, Kazemi K. Novel shape-texture feature extraction for medical X-ray image classification[J]. International Journal of Innovative Computing Information & Control, 2012,8(1):658-676. [8] Li Hailiang, Weng Jian, Shi Yujian, et al. An improved deep learning approach for detection of thyroid papillary cancer in ultrasound images[J]. Scientific Reports, 2018, 8(1):6600-6611. [9] Singh N, Jindal A. Ultra sonogram images for thyroid segmentation and texture classification in diagnosis of malignant (cancerous) or benign (non-cancerous) nodules[J]. J Eng Innov Technol, 2012,2012: 202-206. [10] Liu Tianjiao, Guo Qianqian, Lian Chunfeng, et al. Automated detection and classification of thyroid nodules in ultrasound images using clinical-knowledge-guided convolutional neural networks[J]. Medical Image Analysis, 2019, 58:101555. [11] Horvath E, Majlis S, Rossi R, et al. An ultrasonogram reporting system for thyroid nodules stratifying cancer risk for clinical management[J].The Journal of Clinical Endocrinology & Metabolism, 2009, 94(5):1748-1751. [12] Nibber A, Mistry R, Sooriyamoorthy T, et al. Ultrasound classification of thyroid nodules: A systematic review[J]. Cureus, 2020, 12(3):e7239. [13] Ma Jinlian, Wu Fa, Zhu Jiang, et al. A pre-trained convolutional neural network based method for thyroid nodule diagnosis[J]. Ultrasonics, 2017, 73:221-230. [14] 孔小函, 檀韬, 包凌云,等. 基于卷积神经网络和多信息融合的三维乳腺超声分类方法[J]. 中国生物医学工程学报, 2018, 37(4):414-422. [15] Ren Shaoqing, He Kaiming, Girshick R, et al. Faster R-CNN: Towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2017, 39(6):1137-1149. [16] Li H, Weng J, Shi Y, et al. An improved deep learning approach for detection of thyroid papillary cancer in ultrasound images[J]. Scientific Reports, 2018, 8(1):6600. [17] 宋尚玲, 杨阳, 李夏,等. 基于Faster-RCNN的肺结节检测算法[J]. 中国生物医学工程学报, 2020, 39(2):129-136. [18] Cai Z, Vasconcelos N. Cascade R-CNN: delving into high quality object detection[C] //2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Salt Lake City: IEEE, 2018: 6154-6162. [19] Hu Jie, Li Shen, Sun Gang. Squeeze-and-excitation networks[C] //2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Salt Lake City: IEEE, 2018: 7132-7141. [20] 肖志涛, 刘博文, 耿磊, 等. 基于Reception和Faster R-CNN的肺结节检测方法. [P].CN201910220570.4, 2019. [21] Dhaya R. Deep net model for detection of Covid-19 using radiographs based on ROC analysis[J]. Journal of Innovative Image Processing, 2020, 2(3):135-140. [22] Tajbakhsh N, Shin J, Gurudu S, et al. Convolutional neural networks for medical image analysis: Full training or fine tuning?[J]. IEEE Transactions on Medical Imaging, 2016, 35(5), 1299-1312. [23] Zhou Meng,Wang Rui, Fu Peng, et al. Automatic malignant thyroid nodule recognition in ultrasound images based on deep learning[J]. E3S Web of Conferences, 2020, 185(2):03021. [24] Olaf R, Philipp F, Thomas B. U-Net: Convolutional networks for biomedical image segmentation[J]. Springer International Publishing, 2015, 9351: 234-241. [25] 杨雷, 孙婷婷, 袁延楠, 等. 1995-2010年北京城区甲状腺癌发病趋势及病理特征分析[J]. 中华预防医学杂志, 2013, 47(2): 109-112. [26] Vicario G, Craparotta G, Pistone G. Meta-models in computer experiments: Kriging versus artificial neural networks[J]. Quality and Reliability Engineering International, 2016, 32:2055-2065. [27] 战国科. 基于人工神经网络的图像识别方法研究[D]. 北京: 中国计量科学研究院, 2007. [28] 刘鑫童. 深度卷积神经网络在甲状腺超声图像中的分类研究[D].银川: 宁夏大学, 2018. [29] 柯威. 基于深度卷积神经网络的甲状腺乳头状癌超声图像识别的研究[D].广州: 广东工业大学, 2018. [30] He H, Garcia EA. Learning from imbalanced data [J]. IEEE Transactions on Knowledge and Data Engineering, 2009, 21(9): 1263-1284. [31] Aishwarya KC, Gannamaneni S, Gowda G, et al. TIRADS classification of thyroid nodules: A pictorial review [J] International Journal of Medicine Research, 2019, 4(2):35-40.
