A Risk Stratification Study of Ultrasound Images of Thyroid Nodules Based on Improved DETR

被引:0
|
作者
Le, Zhang [1 ]
Liang, Yue [1 ]
Hu, Xiaokang [1 ]
Qiu, Taorong [1 ]
Xu, Pan [2 ]
机构
[1] Nanchang Univ, Sch Math & Comp Sci, Nanchang, Peoples R China
[2] Nanchang Univ, Affiliated Hosp 1, Nanchang, Peoples R China
关键词
computer-aided diagnosis; C-TIRADS; multi-label object detection; thyroid nodule ultrasound images;
D O I
10.1002/ima.23219
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
The Chinese Thyroid Imaging Reporting and Data System (C-TIRADS) standard is based on the Chinese current medical context. However, at present, there is a lack of C-TIRADS-based automatic computer-aided diagnosis system for thyroid nodule ultrasound images, and the existing algorithms for detecting and recognizing thyroid nodules are basically for the dichotomous classification of benign and malignant. We used the DETR (detection transformer) model as a baseline model and carried out model enhancements to address the shortcomings of unsatisfactory classification accuracy and difficulty in detecting small thyroid nodules in the DETR model. First, to investigate the method of acquiring multi-scale features of thyroid nodule ultrasound images, we choose TResNet-L as the feature extraction network and introduce multi-scale feature information and group convolution, thereby enhancing the model's multi-label classification accuracy. Second, a parallel decoder architecture for multi-label thyroid nodule ultrasound image classification is designed to enhance the learning of correlation between pathological feature class labels, aiming to improve the multi-label classification accuracy of the detection model. Third, the loss function of the detection model is improved. We propose a linear combination of Smooth L1-Loss and CIoU Loss as the model's bounding box loss function and asymmetric loss as the model's multi-label classification loss function, aiming to further improve the detection model's detection accuracy for small thyroid nodules. The experiment results show that the improved DETR model achieves an AP of 92.4% and 81.6% with IoU thresholds of 0.5 and 0.75, respectively.
引用
收藏
页数:13
相关论文
共 50 条
  • [41] Interobserver Agreement in Ultrasound Risk Stratification Systems for Thyroid Nodules on Static Images Versus Cine-Loop Video Sequences
    Schenke, Simone Agnes
    Petersen, Manuela
    Goerges, Rainer
    Ruhlmann, Verena
    Zimny, Michael
    Richter, Johannes-Paul
    Groener, Daniel
    Baumgarten, Justus
    Kreissl, Michael C.
    Stahl, Alexander R.
    Grunert, Michael
    Klemenz, Burkhard
    Veit, Franziska
    Zettinig, Georg
    Seifert, Philipp
    DIAGNOSTICS, 2024, 14 (19)
  • [42] Classification of thyroid nodules using ultrasound images
    Manivannan, T.
    Ayyappan, Nagarajan
    BIOINFORMATION, 2020, 16 (02) : 145 - 148
  • [43] Deep learning on ultrasound images of thyroid nodules
    Sharifi, Yasaman
    Bakhshali, Mohamad Amin
    Dehghani, Toktam
    DanaiAshgzari, Morteza
    Sargolzaei, Mahdi
    Eslami, Saeid
    BIOCYBERNETICS AND BIOMEDICAL ENGINEERING, 2021, 41 (02) : 636 - 655
  • [44] Performance of current ultrasound-based malignancy risk stratification systems for thyroid nodules in patients with follicular neoplasms
    Lin, Yinghe
    Lai, Shuiqing
    Wang, Peiqing
    Li, Jinlian
    Chen, Zhijiang
    Wang, Long
    Guan, Haixia
    Kuang, Jian
    EUROPEAN RADIOLOGY, 2022, 32 (06) : 3617 - 3630
  • [45] Performance of current ultrasound-based malignancy risk stratification systems for thyroid nodules in patients with follicular neoplasms
    Yinghe Lin
    Shuiqing Lai
    Peiqing Wang
    Jinlian Li
    Zhijiang Chen
    Long Wang
    Haixia Guan
    Jian Kuang
    European Radiology, 2022, 32 : 3617 - 3630
  • [46] Automated location of thyroid nodules in ultrasound images with improved YOLOV3 network
    Zhang, Ling
    Zhuang, Yan
    Hua, Zhan
    Han, Lin
    Li, Cheng
    Chen, Ke
    Peng, Yulan
    Lin, Jiangli
    JOURNAL OF X-RAY SCIENCE AND TECHNOLOGY, 2021, 29 (01) : 75 - 90
  • [47] Deep learning based classification of ultrasound images for thyroid nodules: a large scale of pilot study
    Guan, Qing
    Wang, Yunjun
    Du, Jiajun
    Qin, Yu
    Lu, Hongtao
    Xiang, Jun
    Wang, Fen
    ANNALS OF TRANSLATIONAL MEDICINE, 2019, 7 (07)
  • [48] Risk for malignancy of thyroid nodules based on ultrasound imaging characteristics
    Wu, Qi
    Li, Juanjuan
    Wu, Juan
    Li, Xiang
    Chen, Chuang
    Jiang, Nan
    Sun, Shengrong
    INTERNATIONAL JOURNAL OF CLINICAL AND EXPERIMENTAL MEDICINE, 2016, 9 (06): : 11817 - 11823
  • [49] European Thyroid Association Guidelines for Ultrasound Malignancy Risk Stratification of Thyroid Nodules in Adults: The EU-TIRADS
    Russ, Gilles
    Bonnema, Steen J.
    Erdogan, Murat Faik
    Durante, Cosimo
    Ngu, Rose
    Leenhardt, Laurence
    EUROPEAN THYROID JOURNAL, 2017, 6 (05) : 225 - 237
  • [50] Establishment of an Ultrasound Malignancy Risk Stratification Model for Thyroid Nodules Larger Than 4 cm
    Xi, Xuehua
    Wang, Ying
    Gao, Luying
    Jiang, Yuxin
    Liang, Zhiyong
    Ren, Xinyu
    Gao, Qing
    Lai, Xingjian
    Yang, Xiao
    Zhu, Shenling
    Zhao, Ruina
    Zhang, Xiaoyan
    Zhang, Bo
    FRONTIERS IN ONCOLOGY, 2021, 11