DS-AdaptNet: An Efficient Retinal Vessel Segmentation Framework With Adaptive Enhancement and Depthwise Separable Convolutions
Medical image segmentation plays a crucial role in diagnosis and treatment planning, yet faces persistent challenges including limited annotated data, boundary ambiguity, and high computational demands that hinder clinical deployment. This paper presents DS-AdaptNet, an efficient segmentation framew...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/11071315/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849718632534245376 |
|---|---|
| author | Shuting Chen Chengxi Hong Hong Jia |
| author_facet | Shuting Chen Chengxi Hong Hong Jia |
| author_sort | Shuting Chen |
| collection | DOAJ |
| description | Medical image segmentation plays a crucial role in diagnosis and treatment planning, yet faces persistent challenges including limited annotated data, boundary ambiguity, and high computational demands that hinder clinical deployment. This paper presents DS-AdaptNet, an efficient segmentation framework that addresses these challenges through two synergistic technical innovations. First, we introduce a Multi-Dimensional Morphological Perturbation Augmentation Technique (MD-PAT) that generates anatomically plausible variations through topologically-constrained deformation fields, significantly enhancing training data diversity while preserving critical structural properties. Second, we develop a Context-Aware Adaptive Threshold Optimization (CA-ATO) algorithm that dynamically determines optimal thresholds by integrating multi-scale contextual information and uncertainty estimates, substantially improving boundary delineation accuracy and fine structure preservation. These techniques are integrated with an Efficient Depthwise Convolutional Neural Network (ED-CNN) architecture that employs depth-separable convolutions, dramatically reducing computational complexity while maintaining high segmentation accuracy. Our comprehensive experiments on three benchmark retinal vessel segmentation datasets demonstrate that the proposed DS-AdaptNet achieves state-of-the-art performance while maintaining exceptional efficiency. Notably, our method attains a Dice coefficient of 0.8328 on DRIVE, 0.8110 on CHASE_DB1, and 0.8515 on STARE, consistently outperforming existing approaches. Most importantly, DS-AdaptNet achieves these results with only 1.57M parameters and 44.08 GFLOPs—a 94.9% reduction in parameters and 77.2% reduction in computational operations compared to standard U-Net. These efficiency gains enable real-time retinal vessel analysis on standard hardware without specialized acceleration, making DS-AdaptNet particularly suitable for resource-constrained clinical environments and telemedicine applications. The proposed framework establishes a foundation for developing practical computer-aided diagnostic systems that balance accuracy, efficiency, and clinical utility. |
| format | Article |
| id | doaj-art-a7322176686c4bb1a02cb57667dc3160 |
| institution | DOAJ |
| issn | 2169-3536 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-a7322176686c4bb1a02cb57667dc31602025-08-20T03:12:19ZengIEEEIEEE Access2169-35362025-01-011312220712222310.1109/ACCESS.2025.358561111071315DS-AdaptNet: An Efficient Retinal Vessel Segmentation Framework With Adaptive Enhancement and Depthwise Separable ConvolutionsShuting Chen0https://orcid.org/0000-0001-6732-3822Chengxi Hong1https://orcid.org/0009-0007-0540-0678Hong Jia2Chengyi College, Jimei University, Xiamen, ChinaChengyi College, Jimei University, Xiamen, ChinaFujian Key Laboratory of Sensing and Computing for Smart Cities, School of Informatics, Xiamen University, Xiamen, ChinaMedical image segmentation plays a crucial role in diagnosis and treatment planning, yet faces persistent challenges including limited annotated data, boundary ambiguity, and high computational demands that hinder clinical deployment. This paper presents DS-AdaptNet, an efficient segmentation framework that addresses these challenges through two synergistic technical innovations. First, we introduce a Multi-Dimensional Morphological Perturbation Augmentation Technique (MD-PAT) that generates anatomically plausible variations through topologically-constrained deformation fields, significantly enhancing training data diversity while preserving critical structural properties. Second, we develop a Context-Aware Adaptive Threshold Optimization (CA-ATO) algorithm that dynamically determines optimal thresholds by integrating multi-scale contextual information and uncertainty estimates, substantially improving boundary delineation accuracy and fine structure preservation. These techniques are integrated with an Efficient Depthwise Convolutional Neural Network (ED-CNN) architecture that employs depth-separable convolutions, dramatically reducing computational complexity while maintaining high segmentation accuracy. Our comprehensive experiments on three benchmark retinal vessel segmentation datasets demonstrate that the proposed DS-AdaptNet achieves state-of-the-art performance while maintaining exceptional efficiency. Notably, our method attains a Dice coefficient of 0.8328 on DRIVE, 0.8110 on CHASE_DB1, and 0.8515 on STARE, consistently outperforming existing approaches. Most importantly, DS-AdaptNet achieves these results with only 1.57M parameters and 44.08 GFLOPs—a 94.9% reduction in parameters and 77.2% reduction in computational operations compared to standard U-Net. These efficiency gains enable real-time retinal vessel analysis on standard hardware without specialized acceleration, making DS-AdaptNet particularly suitable for resource-constrained clinical environments and telemedicine applications. The proposed framework establishes a foundation for developing practical computer-aided diagnostic systems that balance accuracy, efficiency, and clinical utility.https://ieeexplore.ieee.org/document/11071315/Medical image segmentationretinal vessel segmentationadaptive data augmentationdepth-separable convolutionsadaptive thresholdingmorphology-aware processing |
| spellingShingle | Shuting Chen Chengxi Hong Hong Jia DS-AdaptNet: An Efficient Retinal Vessel Segmentation Framework With Adaptive Enhancement and Depthwise Separable Convolutions IEEE Access Medical image segmentation retinal vessel segmentation adaptive data augmentation depth-separable convolutions adaptive thresholding morphology-aware processing |
| title | DS-AdaptNet: An Efficient Retinal Vessel Segmentation Framework With Adaptive Enhancement and Depthwise Separable Convolutions |
| title_full | DS-AdaptNet: An Efficient Retinal Vessel Segmentation Framework With Adaptive Enhancement and Depthwise Separable Convolutions |
| title_fullStr | DS-AdaptNet: An Efficient Retinal Vessel Segmentation Framework With Adaptive Enhancement and Depthwise Separable Convolutions |
| title_full_unstemmed | DS-AdaptNet: An Efficient Retinal Vessel Segmentation Framework With Adaptive Enhancement and Depthwise Separable Convolutions |
| title_short | DS-AdaptNet: An Efficient Retinal Vessel Segmentation Framework With Adaptive Enhancement and Depthwise Separable Convolutions |
| title_sort | ds adaptnet an efficient retinal vessel segmentation framework with adaptive enhancement and depthwise separable convolutions |
| topic | Medical image segmentation retinal vessel segmentation adaptive data augmentation depth-separable convolutions adaptive thresholding morphology-aware processing |
| url | https://ieeexplore.ieee.org/document/11071315/ |
| work_keys_str_mv | AT shutingchen dsadaptnetanefficientretinalvesselsegmentationframeworkwithadaptiveenhancementanddepthwiseseparableconvolutions AT chengxihong dsadaptnetanefficientretinalvesselsegmentationframeworkwithadaptiveenhancementanddepthwiseseparableconvolutions AT hongjia dsadaptnetanefficientretinalvesselsegmentationframeworkwithadaptiveenhancementanddepthwiseseparableconvolutions |