Multi -task contrastive learning for automatic CT and X-ray diagnosis of COVID-19

PATTERN RECOGNITION

卷:114

Article

Li, Jinpeng^1,2,5  Zhao, Gangming^2,3  Tao, Yaling^1,2  Zhai, Penghua²  Chen, Hao²  He, Huiguang^4,5  Cai, Ting^1,2,5 

[1] Univ Chinese Acad Sci, HwaMei Hosp, 41 Northwest St, Ningbo 315010, Peoples R China

[2] Univ Chinese Acad Sci, Ningbo Inst Life & Hlth Ind, 159 Beijiao St, Ningbo 315000, Peoples R China

[3] Univ Hong Kong, Hong Kong, Peoples R China

[4] Chinese Acad Sci, Inst Automat, 95 Zhongguancun East Rd, Beijing 100190, Peoples R China

[5] Univ Chinese Acad Sci, Beijing, Peoples R China

关键词: Computed tomography;    X-ray;    COVID-19;    Deep learning;    Multi-task learning;    Contrastive learning;   

DOI  :  10.1016/j.patcog.2021.107848

来源: Elsevier

【 摘 要 】

Computed tomography (CT) and X-ray are effective methods for diagnosing COVID-19. Although several studies have demonstrated the potential of deep learning in the automatic diagnosis of COVID-19 using CT and X-ray, the generalization on unseen samples needs to be improved. To tackle this problem, we present the contrastive multi-task convolutional neural network (CMT-CNN), which is composed of two tasks. The main task is to diagnose COVID-19 from other pneumonia and normal control. The auxiliary task is to encourage local aggregation though a contrastive loss: first, each image is transformed by a series of augmentations (Poisson noise, rotation, etc.). Then, the model is optimized to embed represen-tations of a same image similar while different images dissimilar in a latent space. In this way, CMT-CNN is capable of making transformation-invariant predictions and the spread-out properties of data are pre-served. We demonstrate that the apparently simple auxiliary task provides powerful supervisions to en-hance generalization. We conduct experiments on a CT dataset (4,758 samples) and an X-ray dataset (5,821 samples) assembled by open datasets and data collected in our hospital. Experimental results demonstrate that contrastive learning (as plugin module) brings solid accuracy improvement for deep learning models on both CT (5.49%-6.45%) and X-ray (0.96%-2.42%) without requiring additional annota-tions. Our codes are accessible online. (c) 2021 Elsevier Ltd. All rights reserved.

【 授权许可】

Free