Attentional Triple-Encoder Network in Spatiospectral Domains for Medical Image Segmentation
Fuente:
arXiv
Guardado en:
| Autores principales: | , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866912727117594624 |
|---|---|
| author | Qi, Kristin Di, Xinhan |
| author_facet | Qi, Kristin Di, Xinhan |
| contents | Retinal Optical Coherence Tomography (OCT) segmentation is essential for diagnosing pathology. Traditional methods focus on either spatial or spectral domains, overlooking their combined dependencies. We propose a triple-encoder network that integrates CNNs for spatial features, Fast Fourier Convolution (FFC) for spectral features, and attention mechanisms to capture global relationships across both domains. Attention fusion modules integrate convolution and cross-attention to further enhance features. Our method achieves an average Dice score improvement from 0.855 to 0.864, outperforming prior work. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_16389 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Attentional Triple-Encoder Network in Spatiospectral Domains for Medical Image Segmentation Qi, Kristin Di, Xinhan Image and Video Processing Artificial Intelligence Computer Vision and Pattern Recognition Retinal Optical Coherence Tomography (OCT) segmentation is essential for diagnosing pathology. Traditional methods focus on either spatial or spectral domains, overlooking their combined dependencies. We propose a triple-encoder network that integrates CNNs for spatial features, Fast Fourier Convolution (FFC) for spectral features, and attention mechanisms to capture global relationships across both domains. Attention fusion modules integrate convolution and cross-attention to further enhance features. Our method achieves an average Dice score improvement from 0.855 to 0.864, outperforming prior work. |
| title | Attentional Triple-Encoder Network in Spatiospectral Domains for Medical Image Segmentation |
| topic | Image and Video Processing Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.16389 |