TY - GEN
T1 - CFAB
T2 - 14th International Conference on Computer Vision Systems, ICVS 2023
AU - Huang, Jianhua
AU - Huang, Kuan
AU - Xu, Meng
AU - Liu, Feifei
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.
PY - 2023
Y1 - 2023
N2 - Convolutional neural networks (CNNs) may learn spurious correlations between bias features (e.g., background) and labels in image classification. The spuriousness in CNNs usually occurs in building connections between the background of images and labels. Such spurious correlation limits the generalizability of CNNs in classification tasks. Changing backgrounds and foregrounds of original samples can reduce the spuriousness in natural image classification. However, generating annotations for foreground on medical image datasets is time-consuming and labor-intensive. To solve this problem, we propose an online data augmentation method named Combining Foreground And Background (CFAB), which makes CNNs focus on key causal features without foreground annotations and breaks the correlation between backgrounds and labels by changing different backgrounds for one sample. Furthermore, we propose a framework for collaborative augmenting samples using CFAB and training CNNs. Comprehensive experiments indicate that the proposed method weakens the spuriousness, improves the generalizability of the model, and achieves state-of-the-art results in medical ultrasound dataset classification.
AB - Convolutional neural networks (CNNs) may learn spurious correlations between bias features (e.g., background) and labels in image classification. The spuriousness in CNNs usually occurs in building connections between the background of images and labels. Such spurious correlation limits the generalizability of CNNs in classification tasks. Changing backgrounds and foregrounds of original samples can reduce the spuriousness in natural image classification. However, generating annotations for foreground on medical image datasets is time-consuming and labor-intensive. To solve this problem, we propose an online data augmentation method named Combining Foreground And Background (CFAB), which makes CNNs focus on key causal features without foreground annotations and breaks the correlation between backgrounds and labels by changing different backgrounds for one sample. Furthermore, we propose a framework for collaborative augmenting samples using CFAB and training CNNs. Comprehensive experiments indicate that the proposed method weakens the spuriousness, improves the generalizability of the model, and achieves state-of-the-art results in medical ultrasound dataset classification.
KW - Classification
KW - Data Augmentation
KW - Medical Ultrasound Image
KW - Spurious Correlation
UR - http://www.scopus.com/inward/record.url?scp=85174529888&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-44137-0_8
DO - 10.1007/978-3-031-44137-0_8
M3 - Conference contribution
AN - SCOPUS:85174529888
SN - 9783031441363
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 91
EP - 101
BT - Computer Vision Systems - 14th International Conference, ICVS 2023, Proceedings
A2 - Christensen, Henrik I.
A2 - Corke, Peter
A2 - Detry, Renaud
A2 - Weibel, Jean-Baptiste
A2 - Vincze, Markus
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 27 September 2023 through 29 September 2023
ER -