GeoTransfer : Generalizable Few-Shot Multi-View Reconstruction via Transfer Learning

• URL: http://arxiv.org/abs/2408.14724v2
• Date: Sat, 28 Sep 2024 16:47:19 GMT
• Title: GeoTransfer : Generalizable Few-Shot Multi-View Reconstruction via Transfer Learning
• Authors: Shubhendu Jena, Franck Multon, Adnane Boukhayma,
• Abstract summary: We present a novel approach for sparse 3D reconstruction by leveraging the expressive power of Neural Radiance Fields (NeRFs) Our proposed method offers the best of both worlds by transferring the information encoded in NeRF features to derive an accurate occupancy field representation. We evaluate our approach on the DTU dataset and demonstrate state-of-the-art performance in terms of reconstruction accuracy.
• Score: 8.452349885923507
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: This paper presents a novel approach for sparse 3D reconstruction by leveraging the expressive power of Neural Radiance Fields (NeRFs) and fast transfer of their features to learn accurate occupancy fields. Existing 3D reconstruction methods from sparse inputs still struggle with capturing intricate geometric details and can suffer from limitations in handling occluded regions. On the other hand, NeRFs excel in modeling complex scenes but do not offer means to extract meaningful geometry. Our proposed method offers the best of both worlds by transferring the information encoded in NeRF features to derive an accurate occupancy field representation. We utilize a pre-trained, generalizable state-of-the-art NeRF network to capture detailed scene radiance information, and rapidly transfer this knowledge to train a generalizable implicit occupancy network. This process helps in leveraging the knowledge of the scene geometry encoded in the generalizable NeRF prior and refining it to learn occupancy fields, facilitating a more precise generalizable representation of 3D space. The transfer learning approach leads to a dramatic reduction in training time, by orders of magnitude (i.e. from several days to 3.5 hrs), obviating the need to train generalizable sparse surface reconstruction methods from scratch. Additionally, we introduce a novel loss on volumetric rendering weights that helps in the learning of accurate occupancy fields, along with a normal loss that helps in global smoothing of the occupancy fields. We evaluate our approach on the DTU dataset and demonstrate state-of-the-art performance in terms of reconstruction accuracy, especially in challenging scenarios with sparse input data and occluded regions. We furthermore demonstrate the generalization capabilities of our method by showing qualitative results on the Blended MVS dataset without any retraining.

Related papers

• Optimizing 3D Geometry Reconstruction from Implicit Neural Representations [2.3940819037450987]
  Implicit neural representations have emerged as a powerful tool in learning 3D geometry. We present a novel approach that both reduces computational expenses and enhances the capture of fine details.
  arXiv  Detail & Related papers  (2024-10-16T16:36:23Z)
• Enhancing Generalizability of Representation Learning for Data-Efficient 3D Scene Understanding [50.448520056844885]
  We propose a generative Bayesian network to produce diverse synthetic scenes with real-world patterns. A series of experiments robustly display our method's consistent superiority over existing state-of-the-art pre-training approaches.
  arXiv  Detail & Related papers  (2024-06-17T07:43:53Z)
• ReconFusion: 3D Reconstruction with Diffusion Priors [104.73604630145847]
  We present ReconFusion to reconstruct real-world scenes using only a few photos. Our approach leverages a diffusion prior for novel view synthesis, trained on synthetic and multiview datasets. Our method synthesizes realistic geometry and texture in underconstrained regions while preserving the appearance of observed regions.
  arXiv  Detail & Related papers  (2023-12-05T18:59:58Z)
• Leveraging Neural Radiance Fields for Uncertainty-Aware Visual Localization [56.95046107046027]
  We propose to leverage Neural Radiance Fields (NeRF) to generate training samples for scene coordinate regression. Despite NeRF's efficiency in rendering, many of the rendered data are polluted by artifacts or only contain minimal information gain.
  arXiv  Detail & Related papers  (2023-10-10T20:11:13Z)
• Robust Geometry-Preserving Depth Estimation Using Differentiable Rendering [93.94371335579321]
  We propose a learning framework that trains models to predict geometry-preserving depth without requiring extra data or annotations. Comprehensive experiments underscore our framework's superior generalization capabilities. Our innovative loss functions empower the model to autonomously recover domain-specific scale-and-shift coefficients.
  arXiv  Detail & Related papers  (2023-09-18T12:36:39Z)
• Instant Continual Learning of Neural Radiance Fields [78.08008474313809]
  Neural radiance fields (NeRFs) have emerged as an effective method for novel-view synthesis and 3D scene reconstruction. We propose a continual learning framework for training NeRFs that leverages replay-based methods combined with a hybrid explicit-implicit scene representation. Our method outperforms previous methods in reconstruction quality when trained in a continual setting, while having the additional benefit of being an order of magnitude faster.
  arXiv  Detail & Related papers  (2023-09-04T21:01:55Z)
• Semi-signed neural fitting for surface reconstruction from unoriented point clouds [53.379712818791894]
  We propose SSN-Fitting to reconstruct a better signed distance field. SSN-Fitting consists of a semi-signed supervision and a loss-based region sampling strategy. We conduct experiments to demonstrate that SSN-Fitting achieves state-of-the-art performance under different settings.
  arXiv  Detail & Related papers  (2022-06-14T09:40:17Z)
• High-Order Residual Network for Light Field Super-Resolution [39.93400777363467]
  Plenoptic cameras usually sacrifice the spatial resolution of their SAIss to acquire information from different viewpoints. We propose a novel high-order residual network to learn the geometric features hierarchically from the light field for reconstruction. Our approach enables high-quality reconstruction even in challenging regions and outperforms state-of-the-art single image or LF reconstruction methods with both quantitative measurements and visual evaluation.
  arXiv  Detail & Related papers  (2020-03-29T18:06:05Z)
• Deep Non-Line-of-Sight Reconstruction [18.38481917675749]
  In this paper, we employ convolutional feed-forward networks for solving the reconstruction problem efficiently. We devise a tailored autoencoder architecture, trained end-to-end reconstruction maps transient images directly to a depth map representation. We demonstrate that our feed-forward network, even though it is trained solely on synthetic data, generalizes to measured data from SPAD sensors and is able to obtain results that are competitive with model-based reconstruction methods.
  arXiv  Detail & Related papers  (2020-01-24T16:05:50Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.