Edge-Preserving Probabilistic Downsampling for Reliable Medical Segmentation in Resource-Constrained Environments
Downsampling, often necessitated due to resource constraints such as limited memory, development timelines, or the need to accelerate learning from large volumes of unlabeled medical data, results in the loss of fine-grained details, including small objects and thin class boundaries. When applied to...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10857274/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832088029432381440 |
|---|---|
| author | Shahzad Ali Yu Rim Lee Soo Young Park Won Young Tak Soon Ki Jung |
| author_facet | Shahzad Ali Yu Rim Lee Soo Young Park Won Young Tak Soon Ki Jung |
| author_sort | Shahzad Ali |
| collection | DOAJ |
| description | Downsampling, often necessitated due to resource constraints such as limited memory, development timelines, or the need to accelerate learning from large volumes of unlabeled medical data, results in the loss of fine-grained details, including small objects and thin class boundaries. When applied to ground truth (GT) labels without proper care, downsampling can significantly impair the segmentation network’s ability to accurately interpret and predict detailed features. This, in turn, leads to a degraded performance compared to networks trained on higher-resolution labels. This situation exemplifies the trade-off between computational efficiency and segmentation accuracy, with higher downsampling factors further impairing performance. Preserving critical information during label downsampling is particularly crucial for medical image segmentation. This study introduces a novel approach, Edge-Preserving Probabilistic Downsampling (EPD), designed to retain critical details and bridge the performance gap between networks trained on original and downsampled resolutions. It leverages class uncertainty within a local window to produce soft labels, with the window size determining the downsampling factor. This approach enables segmentation networks to maintain high-quality predictions at lower resolutions. Beyond preserving edge details more effectively than conventional nearest-neighbor downsampling, the proposed algorithm, when adapted for images, exhibits notable performance improvements over conventional bilinear interpolation. Experimental results indicate EPD improves Intersection over Union (IoU) for a multi-class abdominal CT dataset by 2.54%, 10.66%, and 16.31% when downsampling to ½, ¼, and ⅛, respectively, compared to conventional interpolation methods. |
| format | Article |
| id | doaj-art-89706417ac7b49c2b77c9dd744aad32f |
| institution | Kabale University |
| issn | 2169-3536 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-89706417ac7b49c2b77c9dd744aad32f2025-02-06T00:00:37ZengIEEEIEEE Access2169-35362025-01-0113216202163410.1109/ACCESS.2025.353628610857274Edge-Preserving Probabilistic Downsampling for Reliable Medical Segmentation in Resource-Constrained EnvironmentsShahzad Ali0https://orcid.org/0000-0002-4949-8335Yu Rim Lee1Soo Young Park2Won Young Tak3Soon Ki Jung4https://orcid.org/0000-0003-0239-6785School of Computer Science and Engineering, Kyungpook National University, Daegu, South KoreaDepartment of Internal Medicine, College of Medicine, Kyungpook National University Hospital, Kyungpook National University, Daegu, South KoreaDepartment of Internal Medicine, College of Medicine, Kyungpook National University Hospital, Kyungpook National University, Daegu, South KoreaDepartment of Internal Medicine, College of Medicine, Kyungpook National University Hospital, Kyungpook National University, Daegu, South KoreaSchool of Computer Science and Engineering, Kyungpook National University, Daegu, South KoreaDownsampling, often necessitated due to resource constraints such as limited memory, development timelines, or the need to accelerate learning from large volumes of unlabeled medical data, results in the loss of fine-grained details, including small objects and thin class boundaries. When applied to ground truth (GT) labels without proper care, downsampling can significantly impair the segmentation network’s ability to accurately interpret and predict detailed features. This, in turn, leads to a degraded performance compared to networks trained on higher-resolution labels. This situation exemplifies the trade-off between computational efficiency and segmentation accuracy, with higher downsampling factors further impairing performance. Preserving critical information during label downsampling is particularly crucial for medical image segmentation. This study introduces a novel approach, Edge-Preserving Probabilistic Downsampling (EPD), designed to retain critical details and bridge the performance gap between networks trained on original and downsampled resolutions. It leverages class uncertainty within a local window to produce soft labels, with the window size determining the downsampling factor. This approach enables segmentation networks to maintain high-quality predictions at lower resolutions. Beyond preserving edge details more effectively than conventional nearest-neighbor downsampling, the proposed algorithm, when adapted for images, exhibits notable performance improvements over conventional bilinear interpolation. Experimental results indicate EPD improves Intersection over Union (IoU) for a multi-class abdominal CT dataset by 2.54%, 10.66%, and 16.31% when downsampling to ½, ¼, and ⅛, respectively, compared to conventional interpolation methods.https://ieeexplore.ieee.org/document/10857274/Downsampling ground truthresizing segmentation labelssoft labelsnearest neighbor interpolationbilinear interpolationsemantic segmentation |
| spellingShingle | Shahzad Ali Yu Rim Lee Soo Young Park Won Young Tak Soon Ki Jung Edge-Preserving Probabilistic Downsampling for Reliable Medical Segmentation in Resource-Constrained Environments IEEE Access Downsampling ground truth resizing segmentation labels soft labels nearest neighbor interpolation bilinear interpolation semantic segmentation |
| title | Edge-Preserving Probabilistic Downsampling for Reliable Medical Segmentation in Resource-Constrained Environments |
| title_full | Edge-Preserving Probabilistic Downsampling for Reliable Medical Segmentation in Resource-Constrained Environments |
| title_fullStr | Edge-Preserving Probabilistic Downsampling for Reliable Medical Segmentation in Resource-Constrained Environments |
| title_full_unstemmed | Edge-Preserving Probabilistic Downsampling for Reliable Medical Segmentation in Resource-Constrained Environments |
| title_short | Edge-Preserving Probabilistic Downsampling for Reliable Medical Segmentation in Resource-Constrained Environments |
| title_sort | edge preserving probabilistic downsampling for reliable medical segmentation in resource constrained environments |
| topic | Downsampling ground truth resizing segmentation labels soft labels nearest neighbor interpolation bilinear interpolation semantic segmentation |
| url | https://ieeexplore.ieee.org/document/10857274/ |
| work_keys_str_mv | AT shahzadali edgepreservingprobabilisticdownsamplingforreliablemedicalsegmentationinresourceconstrainedenvironments AT yurimlee edgepreservingprobabilisticdownsamplingforreliablemedicalsegmentationinresourceconstrainedenvironments AT sooyoungpark edgepreservingprobabilisticdownsamplingforreliablemedicalsegmentationinresourceconstrainedenvironments AT wonyoungtak edgepreservingprobabilisticdownsamplingforreliablemedicalsegmentationinresourceconstrainedenvironments AT soonkijung edgepreservingprobabilisticdownsamplingforreliablemedicalsegmentationinresourceconstrainedenvironments |