TY - JOUR
T1 - A self-supervised feature-standardization-block for cross-domain lung disease classification
AU - Li, Xuechen
AU - Shen, Linlin
AU - Lai, Zhihui
AU - Li, Zhongliang
AU - Yu, Juan
AU - Pu, Zuhui
AU - Mou, Lisha
AU - Cao, Min
AU - Kong, Heng
AU - Li, Yingqi
AU - Dai, Weicai
N1 - Publisher Copyright:
© 2021 Elsevier Inc.
PY - 2022/6
Y1 - 2022/6
N2 - With the advance of deep learning technology, convolutional neural network (CNN) has been wildly used and achieved the state-of-the-art performances in the area of medical image classification. However, most existing medical image classification methods conduct their experiments on only one public dataset. When applying a well-trained model to a different dataset selected from different sources, the model usually shows large performance degradation and needs to be fine-tuned before it can be applied to the new dataset. The goal of this work is trying to solve the cross-domain image classification problem without using data from target domain. In this work, we designed a self-supervised plug-and-play feature-standardization-block (FSB) which consisting of image normalization (INB), contrast enhancement (CEB) and boundary detection blocks (BDB), to extract cross-domain robust feature maps for deep learning framework, and applied the network for chest x-ray-based lung diseases classification. Three classic deep networks, i.e. VGG, Xception and DenseNet and four chest x-ray lung diseases datasets were employed for evaluating the performance. The experimental result showed that when employing feature-standardization-block, all three networks showed better domain adaption performance. The image normalization, contrast enhancement and boundary detection blocks achieved in average 2%, 2% and 5% accuracy improvement, respectively. By combining all three blocks, feature-standardization-block achieved in average 6% accuracy improvement.
AB - With the advance of deep learning technology, convolutional neural network (CNN) has been wildly used and achieved the state-of-the-art performances in the area of medical image classification. However, most existing medical image classification methods conduct their experiments on only one public dataset. When applying a well-trained model to a different dataset selected from different sources, the model usually shows large performance degradation and needs to be fine-tuned before it can be applied to the new dataset. The goal of this work is trying to solve the cross-domain image classification problem without using data from target domain. In this work, we designed a self-supervised plug-and-play feature-standardization-block (FSB) which consisting of image normalization (INB), contrast enhancement (CEB) and boundary detection blocks (BDB), to extract cross-domain robust feature maps for deep learning framework, and applied the network for chest x-ray-based lung diseases classification. Three classic deep networks, i.e. VGG, Xception and DenseNet and four chest x-ray lung diseases datasets were employed for evaluating the performance. The experimental result showed that when employing feature-standardization-block, all three networks showed better domain adaption performance. The image normalization, contrast enhancement and boundary detection blocks achieved in average 2%, 2% and 5% accuracy improvement, respectively. By combining all three blocks, feature-standardization-block achieved in average 6% accuracy improvement.
KW - Chest x-ray
KW - Computer-aided diagnosis
KW - Deep learning
KW - Domain adaption
KW - Lung disease detection
UR - http://www.scopus.com/inward/record.url?scp=85108507619&partnerID=8YFLogxK
U2 - 10.1016/j.ymeth.2021.05.007
DO - 10.1016/j.ymeth.2021.05.007
M3 - Article
C2 - 33992772
AN - SCOPUS:85108507619
SN - 1046-2023
VL - 202
SP - 70
EP - 77
JO - Methods
JF - Methods
ER -