Geometry Reinforced Efficient Attention Tuning Equipped with Normals for Robust Stereo Matching

• URL: http://arxiv.org/abs/2604.09142v1
• Date: Fri, 10 Apr 2026 09:21:51 GMT
• Title: Geometry Reinforced Efficient Attention Tuning Equipped with Normals for Robust Stereo Matching
• Abstract summary: Greaten is a framework that incorporates surface normals as domain-invariant, object-intrinsic, and discriminative geometric cues.<n>It reduces errors by 30% on ETH3D, 8.5% on the non-Lambertian Booster, and 14.1% on KITTI-2015.<n>Greaten-IGEV runs 19.2% faster than GREAT-IGEV and supports high-resolution (3K) inference on Middlebury with disparity ranges up to 768.
• Score: 14.625772079623998
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Despite remarkable advances in image-driven stereo matching over the past decade, Synthetic-to-Realistic Zero-Shot (Syn-to-Real) generalization remains an open challenge. This suboptimal generalization performance mainly stems from cross-domain shifts and ill-posed ambiguities inherent in image textures, particularly in occluded, textureless, repetitive, and non-Lambertian (specular/transparent) regions. To improve Syn-to-Real generalization, we propose GREATEN, a framework that incorporates surface normals as domain-invariant, object-intrinsic, and discriminative geometric cues to compensate for the limitations of image textures. The proposed framework consists of three key components. First, a Gated Contextual-Geometric Fusion (GCGF) module adaptively suppresses unreliable contextual cues in image features and fuses the filtered image features with normal-driven geometric features to construct domain-invariant and discriminative contextual-geometric representations. Second, a Specular-Transparent Augmentation (STA) strategy improves the robustness of GCGF against misleading visual cues in non-Lambertian regions. Third, sparse attention designs preserve the fine-grained global feature extraction capability of GREAT-Stereo for handling occlusion and texture-related ambiguities while substantially reducing computational overhead, including Sparse Spatial (SSA), Sparse Dual-Matching (SDMA), and Simple Volume (SVA) attentions. Trained exclusively on synthetic data such as SceneFlow, GREATEN-IGEV achieves outstanding Syn-to-Real performance. Specifically, it reduces errors by 30% on ETH3D, 8.5% on the non-Lambertian Booster, and 14.1% on KITTI-2015, compared to FoundationStereo, Monster-Stereo, and DEFOM-Stereo, respectively. In addition, GREATEN-IGEV runs 19.2% faster than GREAT-IGEV and supports high-resolution (3K) inference on Middlebury with disparity ranges up to 768.

Related papers

• PhasorNet: Learning Structure from Frequency for Real-Time Stereo Matching [21.209111801863546]
  PhasorNet is a powerful framework that boosts geometric discrimination via frequency-domain cues.<n>With only 5.3M parameters, PhasorNet achieves state-of-the-art performance on the challenging ETH3D benchmark.
  arXiv  Detail & Related papers  (2026-08-30T14:41:36Z)
• EA-LiteUNet: An Edge-Adaptive and Resource-Efficient U-Net for Boundary-Sensitive Dermoscopic Image Segmentation [2.501994007862134]
  We propose EA-LiteUNet, an edge-adaptive and computationally efficient U-Net variant for boundary-sensitive medical image segmentation.<n>The architecture integrates three core mechanisms: boundary-aware representation learning to suppress aliasing and preserve high-frequency structural details.<n>Extensive evaluations across three public dermoscopic datasets demonstrate that EA-LiteUNet consistently achieves superior boundary precision.
  arXiv  Detail & Related papers  (2026-08-16T05:20:39Z)
• HIR-ALIGN: Enhancing Hyperspectral Image Restoration via Diffusion-Based Data Generation [50.46910397782266]
  Hyperspectral image (HSI) restoration is crucial for reliable analysis, as real HSIs suffer from degradations like noise, blur, and resolution loss.<n>We present HIR-ALIGN, a plug-and-play target-adaptive augmentation framework that enhances hyperspectral image restoration by augmenting limited training images with synthetic data that closely matches the target distribution using no extra data.
  arXiv  Detail & Related papers  (2026-05-13T14:14:13Z)
• CFSR: Geometry-Conditioned Shadow Removal via Physical Disentanglement [22.031783191814323]
  CFSR is a multi-modal prior-driven framework that reframes shadow removal as a physics-constrained restoration process.<n> CFSR achieves state-of-the-art performance across multiple challenging benchmarks.
  arXiv  Detail & Related papers  (2026-04-20T09:58:17Z)
• R-PGA: Robust Physical Adversarial Camouflage Generation via Relightable 3D Gaussian Splatting [59.29134814041703]
  Physical adversarial camouflage poses a severe security threat to autonomous driving systems.<n>Current methods fail to generalize across diverse geometric (e.g., viewing configurations) and radiometric variations.<n>We propose the Relightable Physical 3D Gaussian Splatting framework (R-PGA)
  arXiv  Detail & Related papers  (2026-03-27T04:38:04Z)
• 3DGS-HPC: Distractor-free 3D Gaussian Splatting with Hybrid Patch-wise Classification [80.43401018217367]
  3D Gaussian Splatting (3DGS) has demonstrated remarkable performance in novel view synthesis and 3D scene reconstruction.<n>Its quality often degrades in real-world environments due to transient distractors, such as moving objects and varying shadows.<n>Existing methods rely on semantic cues extracted from pre-trained vision models to identify and suppress these distractors.<n>We propose 3DGS- HPC, a framework that circumvents these limitations by combining two complementary principles.
  arXiv  Detail & Related papers  (2026-03-08T11:16:30Z)
• ERGO: Excess-Risk-Guided Optimization for High-Fidelity Monocular 3D Gaussian Splatting [63.138778159026934]
  We propose an adaptive optimization framework guided by excess risk decomposition, termed ERGO.<n> ERGO dynamically estimates the view-specific excess risk and adaptively adjust loss weights during optimization.<n>Experiments on the Google Scanned Objects dataset and the OmniObject3D dataset demonstrate the superiority of ERGO over existing state-of-the-art methods.
  arXiv  Detail & Related papers  (2026-02-10T20:44:43Z)
• STAvatar: Soft Binding and Temporal Density Control for Monocular 3D Head Avatars Reconstruction [18.777422532112105]
  Existing methods based on 3D Gaussian Splatting typically bind Gaussians to mesh triangles and model deformations solely via Linear Blend Skinning.<n>We propose STAvatar, which consists of two key components: (1) a UV-Temporal Soft Binding framework that leverages both image-based and geometric priors to learn per-Gaussian feature offsets within the UV space.<n> STAvatar achieves state-of-the-art reconstruction performance, especially in capturing fine-grained details and reconstructing frequently occluded regions.
  arXiv  Detail & Related papers  (2025-11-25T02:35:00Z)
• 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)
• MCGS: Multiview Consistency Enhancement for Sparse-View 3D Gaussian Radiance Fields [100.90743697473232]
  Radiance fields represented by 3D Gaussians excel at synthesizing novel views, offering both high training efficiency and fast rendering.<n>Existing methods often incorporate depth priors from dense estimation networks but overlook the inherent multi-view consistency in input images.<n>We propose a view synthesis framework based on 3D Gaussian Splatting, enabling scene reconstruction from sparse views.
  arXiv  Detail & Related papers  (2024-10-15T08:39:05Z)
• GS-IR: 3D Gaussian Splatting for Inverse Rendering [71.14234327414086]
  We propose GS-IR, a novel inverse rendering approach based on 3D Gaussian Splatting (GS) We extend GS, a top-performance representation for novel view synthesis, to estimate scene geometry, surface material, and environment illumination from multi-view images captured under unknown lighting conditions. The flexible and expressive GS representation allows us to achieve fast and compact geometry reconstruction, photorealistic novel view synthesis, and effective physically-based rendering.
  arXiv  Detail & Related papers  (2023-11-26T02:35:09Z)
• Hyperspectral Image Super-resolution via Deep Progressive Zero-centric Residual Learning [62.52242684874278]
  Cross-modality distribution of spatial and spectral information makes the problem challenging. We propose a novel textitlightweight deep neural network-based framework, namely PZRes-Net. Our framework learns a high resolution and textitzero-centric residual image, which contains high-frequency spatial details of the scene.
  arXiv  Detail & Related papers  (2020-06-18T06:32:11Z)

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.