TY - GEN
T1 - Degradation-Invariant Enhancement of Fundus Images via Pyramid Constraint Network
AU - Liu, Haofeng
AU - Li, Heng
AU - Fu, Huazhu
AU - Xiao, Ruoxiu
AU - Gao, Yunshu
AU - Hu, Yan
AU - Liu, Jiang
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022
Y1 - 2022
N2 - As an economical and efficient fundus imaging modality, retinal fundus images have been widely adopted in clinical fundus examination. Unfortunately, fundus images often suffer from quality degradation caused by imaging interferences, leading to misdiagnosis. Despite impressive enhancement performances that state-of-the-art methods have achieved, challenges remain in clinical scenarios. For boosting the clinical deployment of fundus image enhancement, this paper proposes the pyramid constraint to develop a degradation-invariant enhancement network (PCE-Net), which mitigates the demand for clinical data and stably enhances unknown data. Firstly, high-quality images are randomly degraded to form sequences of low-quality ones sharing the same content (SeqLCs). Then individual low-quality images are decomposed to Laplacian pyramid features (LPF) as the multi-level input for the enhancement. Subsequently, a feature pyramid constraint (FPC) for the sequence is introduced to enforce the PCE-Net to learn a degradation-invariant model. Extensive experiments have been conducted under the evaluation metrics of enhancement and segmentation. The effectiveness of the PCE-Net was demonstrated in comparison with state-of-the-art methods and the ablation study. The source code of this study is publicly available at https://github.com/HeverLaw/PCENet-Image-Enhancement.
AB - As an economical and efficient fundus imaging modality, retinal fundus images have been widely adopted in clinical fundus examination. Unfortunately, fundus images often suffer from quality degradation caused by imaging interferences, leading to misdiagnosis. Despite impressive enhancement performances that state-of-the-art methods have achieved, challenges remain in clinical scenarios. For boosting the clinical deployment of fundus image enhancement, this paper proposes the pyramid constraint to develop a degradation-invariant enhancement network (PCE-Net), which mitigates the demand for clinical data and stably enhances unknown data. Firstly, high-quality images are randomly degraded to form sequences of low-quality ones sharing the same content (SeqLCs). Then individual low-quality images are decomposed to Laplacian pyramid features (LPF) as the multi-level input for the enhancement. Subsequently, a feature pyramid constraint (FPC) for the sequence is introduced to enforce the PCE-Net to learn a degradation-invariant model. Extensive experiments have been conducted under the evaluation metrics of enhancement and segmentation. The effectiveness of the PCE-Net was demonstrated in comparison with state-of-the-art methods and the ablation study. The source code of this study is publicly available at https://github.com/HeverLaw/PCENet-Image-Enhancement.
KW - Degradation-invariant enhancement
KW - Feature pyramid constraint
KW - Fundus image
KW - Laplacian pyramid feature
UR - http://www.scopus.com/inward/record.url?scp=85139075694&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-16434-7_49
DO - 10.1007/978-3-031-16434-7_49
M3 - Conference contribution
AN - SCOPUS:85139075694
SN - 9783031164330
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 507
EP - 516
BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings
A2 - Wang, Linwei
A2 - Dou, Qi
A2 - Fletcher, P. Thomas
A2 - Speidel, Stefanie
A2 - Li, Shuo
PB - Springer Science and Business Media Deutschland GmbH
T2 - 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022
Y2 - 18 September 2022 through 22 September 2022
ER -