Abstract
We introduce NeSM, a NeRF-based 3D segmentation method, integrating a pre-trained segmentation network's annotation maps as 2D priors through a feature-wise linear modulation (FiLM) layer. By employing this approach, the network can produce high-accuracy segmentation predictions for 3D ultrasound (US) images. NeSM outperforms existing methods, demonstrating higher pixel accuracy, dice coefficient, and mean intersection over union (mIoU) for ultrasound (US) image segmentation.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the 2024 10th International Conference on Applied System Innovation, ICASI 2024 |
| Editors | Shoou-Jinn Chang, Sheng-Joue Young, Artde Donald Kin-Tak Lam, Liang-Wen Ji, Stephen D. Prior |
| Publisher | Institute of Electrical and Electronics Engineers Inc. |
| Pages | 397-399 |
| Number of pages | 3 |
| ISBN (Electronic) | 9798350394924 |
| DOIs | |
| State | Published - 2024 |
| Event | 10th International Conference on Applied System Innovation, ICASI 2024 - Kyoto, Japan Duration: 17 04 2024 → 21 04 2024 |
Publication series
| Name | Proceedings of the 2024 10th International Conference on Applied System Innovation, ICASI 2024 |
|---|
Conference
| Conference | 10th International Conference on Applied System Innovation, ICASI 2024 |
|---|---|
| Country/Territory | Japan |
| City | Kyoto |
| Period | 17/04/24 → 21/04/24 |
Bibliographical note
Publisher Copyright:© 2024 IEEE.
Keywords
- 3D image reconstruction
- 3D semantic segmentation
- Neural radiance field
- implicit neural representation
- medical visualization
- ultrasound
- volume rendering