An Ultrasonic Image Recognition Method for Papillary Thyroid Carcinoma Based on Depth Convolution Neural Network
The ultrasonic image of thyroid papillary carcinoma is characterized by two dimensional gray scale, low resolution, and complicated internal tissue structure. The characteristics of thyroid papillary carcinoma are not obvious and it is difficult to be distinguished. In this paper, the convolution neural network (CNN) theory is introduced for the automatic identification of the ultrasonic image of thyroid papillary carcinoma. Based on the improvement of the Faster RCNN, a detection method for the identification of ultrasonic image features of papillary thyroid carcinoma is proposed, that is, by connecting the fourth and fifth layers of the shared convolution layer in the Faster RCNN, and then normalizing. Secondly, multi-scale ultrasonic images are used in the input. Finally, thyroid papillary carcinoma is classified according to several major characteristics of its ultrasound image.so that the detailed ultrasound image diagnostic report can be received. The experimental results show that the recognition accuracy of the Faster RCNN is higher, the training time is shorter and the efficiency is higher compared with that of the original Faster RCNN.
KeywordsConvolution neural network Thyroid papillary carcinoma Ultrasound images
This work was supported in part by the Degree and Graduate Education Reform Project of Guangdong Province under Grant No. 2016JGXM_MS_26, Foundation of Key Laboratory of Machine Intelligence and Advanced Computing of the Ministry of Education under Grant No. MSC-201706A and Higher Education Quality Project of Guangdong Province.
- 3.Toki, Y., Tanaka, T.: Image feature extraction method with SIFT to diagnose prostate cancer. In: Sice Conference, pp. 2185–2188. IEEE (2012)Google Scholar
- 4.Niwas, S.I., Palanisamy, P., Zhang, W.J., et al.: Log-gabor wavelets based breast carcinoma classification using least square support vector machine. In: IEEE International Conference on Imaging Systems and Techniques, pp. 219–223. IEEE (2011)Google Scholar
- 7.He, K., Zhang, X., Ren, S., et al.: Deep residual learning for image recognition. In: Computer Vision and Pattern Recognition, Las Vegas, Nevada, USA, pp. 770–778 (2016)Google Scholar
- 9.He, K., Zhang, X., Ren, S., et al.: Delving deep into rectifiers: surpassing human-level performance on imagenet classification. In: International Conference on Computer Vision, Santiago, Chile, pp. 1026–1034 (2015)Google Scholar
- 10.Girshick, R., Donahue, J., Darrell, T., et al.: Rich feature hierarchies for accurate object detection and semantic segmentation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Columbus, Ohio, USA, pp. 580–587 (2014)Google Scholar
- 11.Girshick, R.: Fast R-CNN. In: Proceedings of the IEEE International Conference on Computer Vision, Santiago, Chile, pp. 1440–1448 (2015)Google Scholar
- 12.Ren, S., He, K., Girshick, R., et al.: Faster R-CNN: towards real-time object detection with region proposal networks. In: Advances in Neural Information Processing Systems, Montréal, Canada, pp. 91–99 (2015)Google Scholar