Source code for "Quantifying Regional Contributions to Sex Classification from Fundus Photographs via a Two-Stage Attention-Based Deep Learning Approach"
Fuente:
Zenodo
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Recurso digital |
| Veröffentlicht: |
Zenodo
2026
|
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866901605644763136 |
|---|---|
| author | Kim, Joseph Jang, Shina |
| author_facet | Kim, Joseph Jang, Shina |
| contents | <p>Initial public release of source code for the manuscript:</p> <p><strong>Quantifying Regional Contributions to Sex Classification from Fundus Photographs via a Two-Stage Attention-Based Deep Learning Approach</strong></p> <p>This release contains the full pipeline:</p> <ul> <li>U-Net training and inference for vessel segmentation (FIVES dataset)</li> <li>Region-of-interest extraction (vessel, optic disc, macula)</li> <li>Two-stage training (per-branch pre-training + attention fusion)</li> <li>Model evaluation</li> <li>Bootstrap-based statistical analysis with paired 95% CIs and p-values</li> </ul> <p>Per IRB restrictions, pretrained model weights and processed model outputs are not included.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_19935337 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Source code for "Quantifying Regional Contributions to Sex Classification from Fundus Photographs via a Two-Stage Attention-Based Deep Learning Approach" Kim, Joseph Jang, Shina <p>Initial public release of source code for the manuscript:</p> <p><strong>Quantifying Regional Contributions to Sex Classification from Fundus Photographs via a Two-Stage Attention-Based Deep Learning Approach</strong></p> <p>This release contains the full pipeline:</p> <ul> <li>U-Net training and inference for vessel segmentation (FIVES dataset)</li> <li>Region-of-interest extraction (vessel, optic disc, macula)</li> <li>Two-stage training (per-branch pre-training + attention fusion)</li> <li>Model evaluation</li> <li>Bootstrap-based statistical analysis with paired 95% CIs and p-values</li> </ul> <p>Per IRB restrictions, pretrained model weights and processed model outputs are not included.</p> |
| title | Source code for "Quantifying Regional Contributions to Sex Classification from Fundus Photographs via a Two-Stage Attention-Based Deep Learning Approach" |
| url | https://doi.org/10.5281/zenodo.19935337 |