HyperPlanes: Hypernetwork Approach to Rapid NeRF Adaptation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916112846815232 |
|---|---|
| author | Batorski, Paweł Malarz, Dawid Przewięźlikowski, Marcin Mazur, Marcin Tadeja, Sławomir Spurek, Przemysław |
| author_facet | Batorski, Paweł Malarz, Dawid Przewięźlikowski, Marcin Mazur, Marcin Tadeja, Sławomir Spurek, Przemysław |
| contents | Neural radiance fields (NeRFs) are a widely accepted standard for synthesizing new 3D object views from a small number of base images. However, NeRFs have limited generalization properties, which means that we need to use significant computational resources to train individual architectures for each item we want to represent. To address this issue, we propose a few-shot learning approach based on the hypernetwork paradigm that does not require gradient optimization during inference. The hypernetwork gathers information from the training data and generates an update for universal weights. As a result, we have developed an efficient method for generating a high-quality 3D object representation from a small number of images in a single step. This has been confirmed by direct comparison with the state-of-the-art solutions and a comprehensive ablation study. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_01524 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | HyperPlanes: Hypernetwork Approach to Rapid NeRF Adaptation Batorski, Paweł Malarz, Dawid Przewięźlikowski, Marcin Mazur, Marcin Tadeja, Sławomir Spurek, Przemysław Computer Vision and Pattern Recognition Machine Learning Neural radiance fields (NeRFs) are a widely accepted standard for synthesizing new 3D object views from a small number of base images. However, NeRFs have limited generalization properties, which means that we need to use significant computational resources to train individual architectures for each item we want to represent. To address this issue, we propose a few-shot learning approach based on the hypernetwork paradigm that does not require gradient optimization during inference. The hypernetwork gathers information from the training data and generates an update for universal weights. As a result, we have developed an efficient method for generating a high-quality 3D object representation from a small number of images in a single step. This has been confirmed by direct comparison with the state-of-the-art solutions and a comprehensive ablation study. |
| title | HyperPlanes: Hypernetwork Approach to Rapid NeRF Adaptation |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2402.01524 |