TY - GEN
T1 - FD-SDG
T2 - 20th International Conference on Intelligent Computing , ICIC 2024
AU - Li, Boyang
AU - Li, Haojin
AU - Zhang, Yule
AU - Li, Heng
AU - Chen, Jiangyu
AU - Pan, Fuhai
AU - Chen, Jianwen
AU - Liu, Jiang
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.
PY - 2024
Y1 - 2024
N2 - Single-source domain generalization aims to enhance model performance on unseen target domain test sets using only a single source domain dataset, typically by mitigating domain shifts between domains. In retinal vessel segmentation tasks, differences in dataset composition, such as variations in the proportions of different diseases and imaging noise levels, are considered significant sources of domain shift. However, few previous studies have delved into the mechanisms through which this type of domain shift influences model performance. In this study, we hypothesize that disparities in dataset composition could manifest as differences in distribution patterns of frequency domain features, rendering the model susceptible to overfitting specific patterns. Building on this hypothesis, we propose a novel Frequency Dropout based Single Source Domain Generalization (FD-SDG) framework that employs a Frequency Dropout Randomization mechanism to disentangle complex co-adaptive relationships among features from different frequency bands, thereby enhancing the model’s robustness to variable frequency domain noise patterns in the sample space. Additionally, we introduce a Salient Structure Representation Normalization mechanism to align post-perturbation data features in the feature space using invariant anatomical structures. Through comparison experiments and ablation studies conducted on multiple sets of fundus images across-domain experiments, our method achieves state-of-the-art performance, underscoring its high generalizability and robustness.
AB - Single-source domain generalization aims to enhance model performance on unseen target domain test sets using only a single source domain dataset, typically by mitigating domain shifts between domains. In retinal vessel segmentation tasks, differences in dataset composition, such as variations in the proportions of different diseases and imaging noise levels, are considered significant sources of domain shift. However, few previous studies have delved into the mechanisms through which this type of domain shift influences model performance. In this study, we hypothesize that disparities in dataset composition could manifest as differences in distribution patterns of frequency domain features, rendering the model susceptible to overfitting specific patterns. Building on this hypothesis, we propose a novel Frequency Dropout based Single Source Domain Generalization (FD-SDG) framework that employs a Frequency Dropout Randomization mechanism to disentangle complex co-adaptive relationships among features from different frequency bands, thereby enhancing the model’s robustness to variable frequency domain noise patterns in the sample space. Additionally, we introduce a Salient Structure Representation Normalization mechanism to align post-perturbation data features in the feature space using invariant anatomical structures. Through comparison experiments and ablation studies conducted on multiple sets of fundus images across-domain experiments, our method achieves state-of-the-art performance, underscoring its high generalizability and robustness.
KW - domain randomization
KW - frequency dropout
KW - fundus image
KW - single-source domain generalization
KW - vessel segmentation
UR - http://www.scopus.com/inward/record.url?scp=85200944834&partnerID=8YFLogxK
U2 - 10.1007/978-981-97-5689-6_34
DO - 10.1007/978-981-97-5689-6_34
M3 - Conference contribution
AN - SCOPUS:85200944834
SN - 9789819756889
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 393
EP - 404
BT - Advanced Intelligent Computing in Bioinformatics - 20th International Conference, ICIC 2024, Proceedings
A2 - Huang, De-Shuang
A2 - Zhang, Qinhu
A2 - Guo, Jiayang
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 5 August 2024 through 8 August 2024
ER -