Spectral Probing of Feature Upsamplers in 2D-to-3D Scene Reconstruction

• URL: http://arxiv.org/abs/2603.05787v1
• Date: Fri, 06 Mar 2026 00:35:32 GMT
• Title: Spectral Probing of Feature Upsamplers in 2D-to-3D Scene Reconstruction
• Abstract summary: Recent learnable upsampling methods aim to enhance spatial details, but their impact on 3D awareness remains underexplored.<n>Across classical and learnable upsampling methods on CLIP and DINO backbones, we observe three key findings.<n>Results indicate reconstruction quality is more closely related to preserving spectral structure than to enhancing spatial detail.
• Score: 41.21245865872482
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: A typical 2D-to-3D pipeline takes multi-view images as input, where a Vision Foundation Model (VFM) extracts features that are spatially upsampled to dense representations for 3D reconstruction. If dense features across views preserve geometric consistency, differentiable rendering can recover an accurate 3D representation, making the feature upsampler a critical component. Recent learnable upsampling methods mainly aim to enhance spatial details, such as sharper geometry or richer textures, yet their impact on 3D awareness remains underexplored. To address this gap, we introduce a spectral diagnostic framework with six complementary metrics that characterize amplitude redistribution, structural spectral alignment, and directional stability. Across classical interpolation and learnable upsampling methods on CLIP and DINO backbones, we observe three key findings. First, structural spectral consistency (SSC/CSC) is the strongest predictor of NVS quality, whereas High-Frequency Spectral Slope Drift (HFSS) often correlates negatively with reconstruction performance, indicating that emphasizing high-frequency details alone does not necessarily improve 3D reconstruction. Second, geometry and texture respond to different spectral properties: Angular Energy Consistency (ADC) correlates more strongly with geometry-related metrics, while SSC/CSC influence texture fidelity slightly more than geometric accuracy. Third, although learnable upsamplers often produce sharper spatial features, they rarely outperform classical interpolation in reconstruction quality, and their effectiveness depends on the reconstruction model. Overall, our results indicate that reconstruction quality is more closely related to preserving spectral structure than to enhancing spatial detail, highlighting spectral consistency as an important principle for designing upsampling strategies in 2D-to-3D pipelines.

Related papers

• Surfsvr: 2D Surface Priors as 3D Geometric Regularizers for Sparse Voxel Reconstruction [56.51413081953813]
  SurfSVR is a novel sparse voxel reconstruction paradigm that treats 2D surface priors as explicit 3D geometric regularizers.<n>It consistently improves sparse voxel reconstruction across scenes with substantially different visibility and geometry characteristics.
  arXiv  Detail & Related papers  (2026-08-12T11:25:31Z)
• QuerySplat: Decoupling Geometry and Appearance Representations in 3DGS Prediction [25.05567017307621]
  We propose textbfQuerySplat, a feed-forward 3DGS framework driven by geometric priors and explicit appearance decoupling.<n>Experiments demonstrate that QuerySplat mitigates the blurry rendering issues of early query-based models.<n>On the challenging DL3DV benchmark, it achieves state-of-the-art novel view synthesis performance, with average PSNR gains of 2.30 dB and 1.04 dB over the best pose-free and pose-required baselines.
  arXiv  Detail & Related papers  (2026-08-02T12:12:02Z)
• Reliability-Aware 3D Geometric Injection for Universal Person Re-identification [47.874819213762926]
  We propose a Universal Monocular 3D-Enhanced ReID framework driven by a Consistency-Aware Reliability Gate and Dual-Stream Residual Fusion.<n>Our method improves challenging, structure-sensitive scenarios while preserving competitive performance on clean domains.
  arXiv  Detail & Related papers  (2026-07-21T08:54:25Z)
• FLUX3D: High-Fidelity 3D Gaussian Generation with Diffusion-Aligned Sparse Representation [48.928051741750124]
  We propose a scalable image-to-3DGS framework that boosts representation learning and cross-modal alignment during generation.<n>FLUX3D yields substantial improvements in appearance fidelity and significantly outperforms all state-of-the-art (SOTA) methods in generating high-quality 3DGS assets.
  arXiv  Detail & Related papers  (2026-06-23T17:52:21Z)
• Phy-CoSF: Physics-Guided Continuous Spectral Fields Reconstruction and Super-Resolution for Snapshot Compressive Imaging [79.35520607451187]
  We propose Phy-CoSF, which synergizes deep unfolding networks with implicit neural representations.<n>Specifically, we propose a two-phase architecture that bridges discrete-wavelength training with continuous spectral rendering.
  arXiv  Detail & Related papers  (2026-05-13T14:17:02Z)
• High-Fidelity Surface Splatting-Based 3D Reconstruction from Multi-View Images [0.0]
  Implicit Moving Least Squares (IMLS) enables direct conversion of point clouds into signed distance and texture fields.<n>We introduce a compact kernel with local support and greater flexibility, allowing better control over frequency content.<n> Experiments show state-of-the-art performance in both surface reconstruction and rendering, yielding more accurate geometry and sharper visuals.
  arXiv  Detail & Related papers  (2026-05-08T05:23:28Z)
• SpatialFly: Geometry-Guided Representation Alignment for UAV Vision-and-Language Navigation in Urban Environments [49.966170814478915]
  We propose a geometry-guided spatial representation framework for UAV VLN.<n>Running on RGB observations without explicit 3D reconstruction, SpatialFly introduces a geometry-guided 2D representation alignment mechanism.<n> Experimental results show that SpatialFly consistently outperforms state-of-the-art UAV VLN baselines across both seen and unseen environments.
  arXiv  Detail & Related papers  (2026-03-22T03:56:58Z)
• Exploring Spatiotemporal Feature Propagation for Video-Level Compressive Spectral Reconstruction: Dataset, Model and Benchmark [22.5556672954071]
  Spectral Compressive Imaging (SCI) has achieved remarkable success, unlocking significant potential for dynamic spectral vision.<n>Existing reconstruction methods, primarily image-based, suffer from two limitations.<n>The frame-by-frame reconstruction paradigm fails to ensure temporal consistency, which is crucial in the video perception.
  arXiv  Detail & Related papers  (2026-02-28T12:11:13Z)
• SR$^{2}$-Net: A General Plug-and-Play Model for Spectral Refinement in Hyperspectral Image Super-Resolution [3.4888894498274747]
  HSI-SR aims to enhance spatial resolution while preserving spectrally faithful and physically plausible characteristics.<n>These methods often neglect spectral consistency across bands, leading to spurious oscillations and physically implausible artifacts.<n>We propose a lightweight plug-and-play, physically priors Spectral Rectification Super-Resolution Network (SR$2$-Net) to address this issue.
  arXiv  Detail & Related papers  (2026-01-29T07:00:00Z)
• HBSplat: Robust Sparse-View Gaussian Reconstruction with Hybrid-Loss Guided Depth and Bidirectional Warping [11.035994094874141]
  HBSplat is a framework that seamlessly integrates robust structural cues, virtual view constraints, and occluded region completion.<n> HBSplat sets a new state-of-the-art, achieving up to 21.13 dB PSNR and 0.189 LPIPS, while maintaining real-time inference.
  arXiv  Detail & Related papers  (2025-09-29T15:03:31Z)
• Seeing 3D Through 2D Lenses: 3D Few-Shot Class-Incremental Learning via Cross-Modal Geometric Rectification [59.17489431187807]
  We propose a framework that enhances 3D geometric fidelity by leveraging CLIP's hierarchical spatial semantics.<n>Our method significantly improves 3D few-shot class-incremental learning, achieving superior geometric coherence and robustness to texture bias.
  arXiv  Detail & Related papers  (2025-09-18T13:45:08Z)
• Towards Integrating Multi-Spectral Imaging with Gaussian Splatting [8.908394706712395]
  We present a study of how to integrate color (RGB) and multi-spectral imagery into the 3D Gaussian Splatting framework.<n>We suggest integrating multi-spectral data directly into the spherical harmonics color components to compactly model each Gaussian's multi-spectral reflectance.
  arXiv  Detail & Related papers  (2025-08-31T20:53:35Z)
• DCT-Mamba3D: Spectral Decorrelation and Spatial-Spectral Feature Extraction for Hyperspectral Image Classification [38.538268270711534]
  Hyperspectral image classification presents challenges due to spectral redundancy and complex spatial-spectral dependencies.<n>This paper proposes a novel framework, DCT-Mamba3D, for hyperspectral image classification.
  arXiv  Detail & Related papers  (2025-02-04T04:00:08Z)
• GeoGS3D: Single-view 3D Reconstruction via Geometric-aware Diffusion Model and Gaussian Splatting [81.03553265684184]
  We introduce GeoGS3D, a framework for reconstructing detailed 3D objects from single-view images. We propose a novel metric, Gaussian Divergence Significance (GDS), to prune unnecessary operations during optimization. Experiments demonstrate that GeoGS3D generates images with high consistency across views and reconstructs high-quality 3D objects.
  arXiv  Detail & Related papers  (2024-03-15T12:24:36Z)
• SpectralGPT: Spectral Remote Sensing Foundation Model [60.023956954916414]
  A universal RS foundation model, named SpectralGPT, is purpose-built to handle spectral RS images using a novel 3D generative pretrained transformer (GPT) Compared to existing foundation models, SpectralGPT accommodates input images with varying sizes, resolutions, time series, and regions in a progressive training fashion, enabling full utilization of extensive RS big data. Our evaluation highlights significant performance improvements with pretrained SpectralGPT models, signifying substantial potential in advancing spectral RS big data applications within the field of geoscience.
  arXiv  Detail & Related papers  (2023-11-13T07:09:30Z)

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.