Recollection from Pensieve: Novel View Synthesis via Learning from Uncalibrated Videos

• URL: http://arxiv.org/abs/2505.13440v1
• Date: Mon, 19 May 2025 17:59:05 GMT
• Title: Recollection from Pensieve: Novel View Synthesis via Learning from Uncalibrated Videos
• Authors: Ruoyu Wang, Yi Ma, Shenghua Gao,
• Abstract summary: We propose a novel two-stage strategy to train a view synthesis model from only raw video frames or multi-view images.<n>In the first stage, we learn to reconstruct the scene implicitly in a latent space without relying on any explicit 3D representation.<n>The learned latent camera and implicit scene representation have a large gap compared with the real 3D world.
• Score: 36.49978976710115
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Currently almost all state-of-the-art novel view synthesis and reconstruction models rely on calibrated cameras or additional geometric priors for training. These prerequisites significantly limit their applicability to massive uncalibrated data. To alleviate this requirement and unlock the potential for self-supervised training on large-scale uncalibrated videos, we propose a novel two-stage strategy to train a view synthesis model from only raw video frames or multi-view images, without providing camera parameters or other priors. In the first stage, we learn to reconstruct the scene implicitly in a latent space without relying on any explicit 3D representation. Specifically, we predict per-frame latent camera and scene context features, and employ a view synthesis model as a proxy for explicit rendering. This pretraining stage substantially reduces the optimization complexity and encourages the network to learn the underlying 3D consistency in a self-supervised manner. The learned latent camera and implicit scene representation have a large gap compared with the real 3D world. To reduce this gap, we introduce the second stage training by explicitly predicting 3D Gaussian primitives. We additionally apply explicit Gaussian Splatting rendering loss and depth projection loss to align the learned latent representations with physically grounded 3D geometry. In this way, Stage 1 provides a strong initialization and Stage 2 enforces 3D consistency - the two stages are complementary and mutually beneficial. Extensive experiments demonstrate the effectiveness of our approach, achieving high-quality novel view synthesis and accurate camera pose estimation, compared to methods that employ supervision with calibration, pose, or depth information. The code is available at https://github.com/Dwawayu/Pensieve.

Related papers

• Fast Learning of Non-Cooperative Spacecraft 3D Models through Primitive Initialization [3.686808512438363]
  This work contributes a Convolutional Neural Network (CNN) based primitive initializer for 3DGS using monocular images.<n>A CNN takes a single image as input and outputs a coarse 3D model represented as an assembly of primitives, along with the target's pose relative to the camera.<n>This work performs a comparison between these variants, evaluating their effectiveness for downstream 3DGS training under noisy or implicit pose estimates.
  arXiv  Detail & Related papers  (2025-07-25T17:43:29Z)
• Gaussian2Scene: 3D Scene Representation Learning via Self-supervised Learning with 3D Gaussian Splatting [6.678115792482272]
  Self-supervised learning (SSL) for point cloud pre-training has become a cornerstone for many 3D vision tasks.<n>We propose a novel scene-level SSL framework that leverages the efficiency and explicit nature of 3D Gaussian Splatting (3DGS) for pre-training.
  arXiv  Detail & Related papers  (2025-06-10T13:19:21Z)
• No Pose, No Problem: Surprisingly Simple 3D Gaussian Splats from Sparse Unposed Images [100.80376573969045]
  NoPoSplat is a feed-forward model capable of reconstructing 3D scenes parameterized by 3D Gaussians from multi-view images. Our model achieves real-time 3D Gaussian reconstruction during inference. This work makes significant advances in pose-free generalizable 3D reconstruction and demonstrates its applicability to real-world scenarios.
  arXiv  Detail & Related papers  (2024-10-31T17:58:22Z)
• 3D-free meets 3D priors: Novel View Synthesis from a Single Image with Pretrained Diffusion Guidance [61.06034736050515]
  We introduce a method capable of generating camera-controlled viewpoints from a single input image.<n>Our method excels in handling complex and diverse scenes without extensive training or additional 3D and multiview data.
  arXiv  Detail & Related papers  (2024-08-12T13:53:40Z)
• DistillNeRF: Perceiving 3D Scenes from Single-Glance Images by Distilling Neural Fields and Foundation Model Features [65.8738034806085]
  DistillNeRF is a self-supervised learning framework for understanding 3D environments in autonomous driving scenes. Our method is a generalizable feedforward model that predicts a rich neural scene representation from sparse, single-frame multi-view camera inputs.
  arXiv  Detail & Related papers  (2024-06-17T21:15:13Z)
• Denoising Diffusion via Image-Based Rendering [54.20828696348574]
  We introduce the first diffusion model able to perform fast, detailed reconstruction and generation of real-world 3D scenes. First, we introduce a new neural scene representation, IB-planes, that can efficiently and accurately represent large 3D scenes. Second, we propose a denoising-diffusion framework to learn a prior over this novel 3D scene representation, using only 2D images.
  arXiv  Detail & Related papers  (2024-02-05T19:00:45Z)
• COLMAP-Free 3D Gaussian Splatting [88.420322646756]
  We propose a novel method to perform novel view synthesis without any SfM preprocessing. We process the input frames in a sequential manner and progressively grow the 3D Gaussians set by taking one input frame at a time. Our method significantly improves over previous approaches in view synthesis and camera pose estimation under large motion changes.
  arXiv  Detail & Related papers  (2023-12-12T18:39:52Z)
• PonderV2: Pave the Way for 3D Foundation Model with A Universal Pre-training Paradigm [111.16358607889609]
  We introduce a novel universal 3D pre-training framework designed to facilitate the acquisition of efficient 3D representation.<n>For the first time, PonderV2 achieves state-of-the-art performance on 11 indoor and outdoor benchmarks, implying its effectiveness.
  arXiv  Detail & Related papers  (2023-10-12T17:59:57Z)
• DreamSparse: Escaping from Plato's Cave with 2D Frozen Diffusion Model Given Sparse Views [20.685453627120832]
  Existing methods often struggle with producing high-quality results or necessitate per-object optimization in such few-view settings. DreamSparse is capable of synthesizing high-quality novel views for both object and scene-level images.
  arXiv  Detail & Related papers  (2023-06-06T05:26:26Z)
• RUST: Latent Neural Scene Representations from Unposed Imagery [21.433079925439234]
  Inferring structure of 3D scenes from 2D observations is a fundamental challenge in computer vision. Recent popularized approaches based on neural scene representations have achieved tremendous impact. RUST (Really Unposed Scene representation Transformer) is a pose-free approach to novel view trained on RGB images alone.
  arXiv  Detail & Related papers  (2022-11-25T18:59:10Z)

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.