Pulmonary Lesions Automatic Detection based on Image Processing Technology

10.6919/ICJE.202205_8(5).0013

Sijia Zhang；Zhen Chen；Mengfei Feng

Double Complex Wavelet Transform ； Gray Linear Transformation ； U-net ； Histogram Comparison ； Automatic Detection of Lesions

International Core Journal of Engineering

8卷5期（2022 / 05 / 01）

100 - 108

英文

With the rapid development of medical imaging technology, all kinds of digital image began to widely used in clinical diagnosis, based on the technology of machine automatically detect lesions arises at the historic moment. The cell images of lung section were observed and studied. Firstly, the lung cells to be diagnosed were collected, and the image was denoised by the double number complex wavelet transform to improve the ability of image edge detection. Secondly, gray linear transformation enhancement technology is used to enhance the denoised image, and the image becomes clearer by increasing the contrast.After image pretreatment, since the 2D U-Net image segmentation technology is relatively rough to display the features of lung cells, the improved 3D U-Net image segmentation technology is used to process the lung cell images. In the same environment, the histogram of lung cell images and lung cancer cell images in the hospital are compared and analyzed. The predicted image of lesion distribution was obtained. Through a large number of experiments conducted by the machine, it was found that the results of automatic detection by the machine were similar to the results of the doctor's manual labeling of lesions, so the automatic detection of lesions could be well completed.

1. Li X, Gao S. Trends of incidence, morbidity and mortality of lung cancer among Chinese residents from 1990 to 2019 [J]. Chinese Journal of Chronic Disease Prevention and Control. 201,29(11):821-826.
2. Ji Cong, Yu Changyuan, Chen Zhihua. [J]. Chinese journal of oncology, 2014,41(20):1333-1336. (in Chinese).
3. Yi Li qun. Denoising method of dynamic PET image based on graph wavelet transform [D]. Hubei: Wuhan University of Science and Technology,2022.
4. Zuo Wentao, Chen Jiayi. Double tree complex wavelet and bivariate threshold model for image denoising [J]. Laboratory Research and Exploration, 201,40(10):29-33.
5. Wan Liyong, Chen Jiayi. Image filtering based on double tree complex Wavelet transform and bilateral filtering [J]. Journal of Central China Normal University (Natural Science Edition), 201,55(06):1030-1036.
6. Zhou Tao, DONG Yali, Huo Bingqiang, Liu Shan, Ma Zongjun. Journal of Image and Graphics, 201, 26 (09):2058-2077.
7. Lin Zhijie, Zheng Qiulan, Liang Yong, Xing Wei. Medical image Segmentation Method based on Inner roll U-NET Model [J]. Computer Engineering :1-9.
8. Li Taohua. Research on CT image segmentation of liver and tumor based on U-NET network [D]. Shaanxi Province: Xi 'an University of Science and Technology,2022.
9. Li Guyue. Research on image segmentation algorithm of pulmonary nodules based on U-NET Network [D]. Heilongjiang Province: Harbin Institute of Technology, 2021.
10. Li Zhaoyan. Optimization and implementation of medical image segmentation algorithm based on deep learning [D]. Sichuan Province: University of Electronic Science and Technology of China, 2020.