Robustly detecting mpox and non-mpox using a deep learning framework based on image inpainting.
Yujun Cao, Yubiao Yue, Xiaoming Ma, Di Liu, Rongkai Ni, Haihua Liang, Zhenzhang Li
Author Information
Yujun Cao: Department of Basic Courses, Guangzhou Maritime University, Guangzhou, 510725, China.
Yubiao Yue: School of Biomedical Engineering, The Second Affiliated Hospital of Guangzhou Medical University, Guangzhou Medical University, Guangzhou, 511436, China.
Xiaoming Ma: Department of Basic Courses, Guangzhou Maritime University, Guangzhou, 510725, China.
Di Liu: Department of Basic Courses, Guangzhou Maritime University, Guangzhou, 510725, China.
Rongkai Ni: School of Biomedical Engineering, The Second Affiliated Hospital of Guangzhou Medical University, Guangzhou Medical University, Guangzhou, 511436, China.
Haihua Liang: School of Mathematics and Systems Science, Guangdong Polytechnic Normal University, No. 293, Zhongshan Avenue West, Tianhe District, Guangzhou, 510665, China.
Zhenzhang Li: School of Mathematics and Systems Science, Guangdong Polytechnic Normal University, No. 293, Zhongshan Avenue West, Tianhe District, Guangzhou, 510665, China. zhenzhangli@gpnu.edu.cn.
Due to the lack of efficient mpox diagnostic technology, mpox cases continue to increase. Recently, the great potential of deep learning models in detecting mpox and non-mpox has been proven. However, existing methods are susceptible to interference from various noises in real-world settings, require diverse non-mpox images, and fail to detect abnormal input, which makes them unsuitable for practical deployment and application. To address these challenges, we proposed a novel strategy based on image inpainting called "Mask, Inpainting, and Measure" (MIM). In MIM's pipeline, a generative adversarial network learns feature representations of mpox images by inpainting the masked mpox images. On this basis, MIM measure the similarity between the inpainted image and the original image to detect mpox and non-mpox. Compared with multi-class classification models, MIM can handle unknown categories and abnormal inputs more effectively. We used the recognized mpox dataset (MSLD) and a dataset containing 18 categories of non-mpoxskin diseases to verify the effectiveness and robustness of MIM. Experimental results show that the average AUROC of MIM achieves 0.8237. In addition, external clinical testing further demonstrates the robustness of MIM. Importantly, we developed a free smartphone app to help the public and healthcare professionals detect mpox more conveniently.
BMC Infect Dis. 2023 Jun 27;23(1):438
[PMID: 37370031]
Grants
2022A1515011044/the NSF of Guangdong Province
2022A1515011044/the NSF of Guangdong Province
2021ZDJS028/the project of promoting research capabilities for key constructed disciplines in Guangdong Province
2021ZDJS028/the project of promoting research capabilities for key constructed disciplines in Guangdong Province
20230233/Scientific research project of Guangzhou Education Bureau
20230233/Scientific research project of Guangzhou Education Bureau
2301002693/Youth innovation talent project of guangdong provincial universities
2301002693/Youth innovation talent project of guangdong provincial universities
22GPNUZDJS31/the scientific research capacity improvement project of the doctoral program construction unit of Guangdong Polytechnic Normal University in 2022
22GPNUZDJS31/the scientific research capacity improvement project of the doctoral program construction unit of Guangdong Polytechnic Normal University in 2022
2019KQNCX067/the young creative talents of department education of Guangdong
2019KQNCX067/the young creative talents of department education of Guangdong
