HyVGGT-VO: Tightly Coupled Hybrid Dense Visual Odometry with Feed-Forward Models Junxiang Pan, Lipu Zhou, Baojie Chen tl;dr: OV²SLAM+VGGT-SLAM arxiv.org/abs/2604.02107

Reliev3R: Relieving Feed-forward Reconstruction from Multi-View Geometric Annotations Youyu Chen, Junjun Jiang, Yueru Luo, Kui Jiang, Xianming Liu, Xu Yan, Dave Zhenyu Chen tl;dr: pseudo relative depths and image correspondences as weak supervion arxiv.org/abs/2604.00548

DirectFisheye-GS: Enabling Native Fisheye Input in Gaussian Splatting with Cross-View Joint Optimization Zhengxian Yang, Fei Xie, Xutao Xue, Rui Zhang, Taicheng Huang, Yang Liu, Mengqi Ji, Tao Yu arxiv.org/abs/2604.00648

Diff3R: Feed-forward 3D Gaussian Splatting with Uncertainty-aware Differentiable Optimization Yueh-Cheng Liu, Jozef Hladký, @niessner.bsky.social, @adai.bsky.social tl;dr: implicit differentiation+PCG solver from 3DGS-LM->differentiable 3DGS optimization layer arxiv.org/abs/2604.01030

Implicit Primal-Dual Interior-Point Methods for Quadratic Programming Jon Arrizabalaga, Zachary Manchester tl;dr: KKT conditions->auxiliary variable and softplus retraction map->spectrally bounded KKT system arxiv.org/abs/2604.00364

Compact Keyframe-Optimized Multi-Agent Gaussian Splatting SLAM Monica M.Q. Li, Pierre-Yves Lajoie, Jialiang Liu, @beltrame.mistlab.ca tl;dr: in title arxiv.org/abs/2604.00804

PanoAir: A Panoramic Visual-Inertial SLAM with Cross-Time Real-World UAV Dataset Yiyang Wu, Xiaohu Zhang, Yanjin Du, Tongsu Zhang, Chujun Li, Siyang Chen, Guoyi Zhang, Xiangpeng Xu tl;dr: in title arxiv.org/abs/2604.00852

Neural Harmonic Textures for High-Quality Primitive Based Neural Reconstruction @jcondor.bsky.social, Nicolas Moenne-Loccoz, @merl.in, Piotr Didyk, Zan Gojcic, Qi Wu tl;dr: anchor latent feature vectors on a virtual scaffold surrounding each primitive arxiv.org/abs/2604.01204

Fisheye3R: Adapting Unified 3D Feed-Forward Foundation Models to Fisheye Lenses Ruxiao Duan, Erin Hong, Dongxu Zhao, Eric Turner, Alex Wong, Yunwen Zhou tl;dr: in title arxiv.org/abs/2603.28896

Hierarchical Visual Relocalization with Nearest View Synthesis from Feature Gaussian Splatting Huaqi Tao, Bingxi Liu, Guangcheng Chen, Fulin Tang, Li He, Hong Zhang tl;dr: iterative retrieval-verification-rendering; rendered feature+JamMa->matching arxiv.org/abs/2603.29185

StereoVGGT: A Training-Free Visual Geometry Transformer for Stereo Vision Ziyang Chen, Yansong Qu, You Shen, Xuan Cheng, Liujuan Cao tl;dr: VGGT+DINOv2+MoGe-2 arxiv.org/abs/2603.29368

AA-Splat: Anti-Aliased Feed-forward Gaussian Splatting Taewoo Suh, Sungpyo Kim, Jongmin Park, Munchurl Kim tl;dr: DepthSplat+Mip-Splatting+SurfSplat arxiv.org/abs/2603.29394

Unblur-SLAM: Dense Neural SLAM for Blurry Inputs Qi Zhang, Denis Rozumny, Francesco Girlanda, Sezer Karaoglu, @marcpollefeys.bsky.social, Theo Gevers, @martin-r-oswald.bsky.social tl;dr: different tracking and mapping modules are used depending on blur levels arxiv.org/abs/2603.26810

SceneExpander: Expanding 3D Scenes with Free-Form Inserted Views Zijian He, enjie Liu, Yihao Wang, Weizhi Zhong, Huan Yuan, Kun Gai, Guangrun Wang, Guanbin Li tl;dr: reconstructed scene+inserted views->test-time adaptation with distillation->expanded scene arxiv.org/abs/2603.27084

MultiLoc: Multi-view Guided Relative Pose Regression for Fast and Robust Visual Re-Localization Nobel Dang, Bing Li tl;dr: pose-guided RPR arxiv.org/abs/2603.27170

Complet4R: Geometric Complete 4D Reconstruction Weibang Wang, Kenan Li, Zhuoguang Chen, Yijun Yuan, Hang Zhao tl;dr: 4D VGGT? arxiv.org/abs/2603.27300

From None to All: Self-Supervised 3D Reconstruction via Novel View Synthesis Ranran Huang, Weixun Luo, Ye Mao, Krystian Mikolajczyk tl;dr: SPFSplatV2 variant arxiv.org/abs/2603.27455

MV-RoMa: From Pairwise Matching into Multi-View Track Reconstruction Jongmin Lee, Seungyeop Kang, Sungjoo Yoo tl;dr: pairwise matches->sampling->initial feature tracks->cross-attention->refined multi-view dense matches arxiv.org/abs/2603.27542

TerraSky3D: Multi-View Reconstructions of European Landmarks in 4K Mattia D'Urso, Yuxi Hu, Christian Sormann, Mattia Rossi, Friedrich Fraundorfer tl;dr: new 3D dataset arxiv.org/abs/2603.28287

Fus3D: Decoding Consolidated 3D Geometry from Feed-forward Geometry Transformer Latents Laura Fink, @linusfranke.bsky.social, @grgkopanas.bsky.social, Marc Stamminger, @peterhedman.bsky.social arxiv.org/abs/2603.25827

Scene Grounding In the Wild Tamir Cohen, Leo Segre, Shay Shomer-Chai, Shai Avidan, Hadar Averbuch-Elor tl;dr: use semantics to bridge the domain gaps between real-world images and pseudo-synthetic renderings arxiv.org/abs/2603.26584

DFM-VLA: Iterative Action Refinement for Robot Manipulation via Discrete Flow Matching Jiayi Chen, Wenxuan Song, Shuai Chen, Jingbo Wang, Zhijun Li, Haoang Li tl;dr: iterative full-sequence action refinement through token-level velocity modeling arxiv.org/abs/2603.26320

VLA-OPD: Bridging Offline SFT and Online RL for Vision-Language-Action Models via On-Policy Distillation Zhide Zhong, Haodong Yan, Junfeng Li, Junjie He, Tianran Zhang, Haoang Li tl;dr: Reverse-KL distillation objective for VLA models arxiv.org/abs/2603.26666

GaussFusion: Improving 3D Reconstruction in the Wild with A Geometry-Informed Video Generator @liyuanzzz.bsky.social, Manjunath Narayana, Michal Stary, Will Hutchcroft, Gordon Wetzstein, @ir0armeni.bsky.social arxiv.org/abs/2603.25053

AirSplat: Alignment and Rating for Robust Feed-Forward 3D Gaussian Splatting Minh-Quan Viet Bui, Jaeho Moon, Munchurl Kim tl;dr: FM->FF 3DGS->NVS arxiv.org/abs/2603.25129

ViewSplat: View-Adaptive Dynamic Gaussian Splatting for Feed-Forward Synthesis Moonyeon Jeong, Seunggi Min, Suhyeon Lee, Hongje Seong tl;dr: use MLP to compensate Gaussians arxiv.org/abs/2603.25265

HeSS: Head Sensitivity Score for Sparsity Redistribution in VGGT Yongsung Kim, Wooseok Song, Jaihyun Lew, Hun Hwangbo, Jaehoon Lee, Sungroh Yoon tl;dr: self-attention->Hessian of camera pose & point cloud errors->sensitivity->adaptive sparsification arxiv.org/abs/2603.25336

DC-Reg: Globally Optimal Point Cloud Registration via Tight Bounding with Difference of Convex Programming Wei Lian, Fei Ma, Hang Pan, Zhesen Cui, Wangmeng Zuo tl;dr: Difference of Convex decomposition->tight lower bound of BnB arxiv.org/abs/2603.25442

Fast-dVLA: Accelerating Discrete Diffusion VLA to Real-Time Performance Wenxuan Song, Jiayi Chen, Shuai Chen, Jingbo Wang, Pengxiang Ding, Han Zhao, Yikai Qin, Xinhu Zheng, Donglin Wang, Yan Wang, Haoang Li arxiv.org/abs/2603.25661

Stochastic Ray Tracing for the Reconstruction of 3D Gaussian Splatting Peiyu Xu, Xin Sun, Krishna Mullia, Raymond Fei, @iliyangeo.bsky.social, Shuang Zhao tl;dr: unbiased Monte Carlo estimator for pixel-color gradients arxiv.org/abs/2603.23637

AdvSplat: Adversarial Attacks on Feed-Forward Gaussian Splatting Models Yiran Qiao, Yiren Lu, Yunlai Zhou, Rui Yang, Linlin Hou, Yu Yin, Jing Ma tl;dr: black-box attacks via optimizing perturbations through repeated queries to the feed-forward 3DGS arxiv.org/abs/2603.23686

FilterGS: Traversal-Free Parallel Filtering and Adaptive Shrinking for Large-Scale LoD 3D Gaussian Splatting Yixian Wang, Haolin Yu, Jiadong Tang, Yu Gao, Xihan Wang, Yufeng Yue, Yi Yang tl;dr: no level-by-level traversal; new GTC metric->shrinking arxiv.org/abs/2603.23891

CAM3R: Camera-Agnostic Model for 3D Reconstruction Namitha Guruprasad, Abhay Yadav, Cheng Peng, Rama Chellappa tl;dr: DUSt3R/VGGT handle wide-angle camera models arxiv.org/abs/2603.22631

Predictive Photometric Uncertainty in Gaussian Splatting for Novel View Synthesis Chamuditha Jayanga Galappaththige, Thomas Gottwald, Peter Stehr, Edgar Heinert, Niko Sünderhauf, Dimity Miller, Matthias Rottmann arxiv.org/abs/2603.22786

UniQueR: Unified Query-based Feedforward 3D Reconstruction Chensheng Peng, Quentin Herau, Jiezhi Yang, Yichen Xie, Yihan Hu, Wenzhao Zheng, Matthew Strong, Masayoshi Tomizuka, Wei Zhan tl;dr: 3D queries->VGGT arxiv.org/abs/2603.22851

Drop-In Perceptual Optimization for 3D Gaussian Splatting Ezgi Ozyilkan, Zhiqi Chen, Oren Rippel, Jona Ballé, Kedar Tatwawadi tl;dr: study L1+SSIM / L1+L2+MS-SSIM+LPIPS / Wasserstein Distortion (WD), and regularized version of WD is the winner arxiv.org/abs/2603.23297

Benchmarking Efficient & Effective Camera Pose Estimation Strategies for Novel View Synthesis Jhacson Meza, @martin-r-oswald.bsky.social, @sattlertorsten.bsky.social arxiv.org/abs/2603.20428

RayMap3R: Inference-Time RayMap for Dynamic 3D Reconstruction Feiran Wang, Zezhou Shang, Gaowen Liu, Yan Yan tl;dr: image and RayMap features->main branch; camera-ray tokens->RayMap branch; depth discrepancy->staticness weights->memory updates arxiv.org/abs/2603.20588

The Role and Relationship of Initialization and Densification in 3D Gaussian Splatting Ivan Desiatov, @sattlertorsten.bsky.social tl;dr: no clear winner on densification; dense initialization has minor improvement; densification matters more? arxiv.org/abs/2603.20714

PlanaReLoc: Camera Relocalization in 3D Planar Primitives via Region-Based Structure Matching Hanqiao Ye, Yuzhou Liu, Yangdong Liu, Shuhan Shen tl;dr: associate planar primitives across query image and the map within a learned unified embedding space arxiv.org/abs/2603.20818

SGAD-SLAM: Splatting Gaussians at Adjusted Depth for Better Radiance Fields in RGBD SLAM Pengchong Hu, Zhizhong Han tl;dr: pixel-aligned Gaussians, but allow Gaussians to move along their rays arxiv.org/abs/2603.21055

Two Experts Are Better Than One Generalist: Decoupling Geometry and Appearance for Feed-Forward 3D Gaussian Splatting Hwasik Jeong, Seungryong Lee, Gyeongjin Kang, Seungkwon Yang, Xiangyu Sun, Seungtae Nam, Eunbyung Park tl;dr: in title arxiv.org/abs/2603.21064

F4Splat: Feed-Forward Predictive Densification for Feed-Forward 3D Gaussian Splatting Injae Kim, Chaehyeon Kim, Minseong Bae, Minseok Joo, Hyunwoo J. Kim tl;dr: learn densification score map to adaptively distribute Gaussians arxiv.org/abs/2603.21304

FluidGaussian: Propagating Simulation-Based Uncertainty Toward Functionally-Intelligent 3D Reconstruction Yuqiu Liu, Jialin Song, Marissa Ramirez de Chanlatte, Rochishnu Chowdhury, Rushil Paresh Desai, @wuyangchen.bsky.social, Daniel Martin, Michael Mahoney arxiv.org/abs/2603.21356

PAS3R: Pose-Adaptive Streaming 3D Reconstruction for Long Video Sequences Lanbo Xu, Liang Guo, Caigui Jiang, Cheng Wang tl;dr: inter-frame camera motion and scene frequency cues->frame importance->learning rate adjustment arxiv.org/abs/2603.21436

EpiMask: Leveraging Epipolar Distance Based Masks in Cross-Attention for Satellite Image Matching Rahul Deshmukh, Aditya Chauhan, Avinash Kak tl;dr: use epipolar distance to mask cross-attention in LoFTR arxiv.org/abs/2603.21463

Image-Conditioned Adaptive Parameter Tuning for Visual Odometry Frontends Simone Nascivera, Leonard Bauersfeld, Jeff Delaune, @davidescaramuzza.bsky.social arxiv.org/abs/2603.21785

Repurposing Geometric Foundation Models for Multi-view Diffusion Wooseok Jang, Seonghu Jeon, Jisang Han, Jinhyeok Choi, Minkyung Kwon, Seungryong Kim, @saining.bsky.social, Sainan Liu tl;dr: feature space of VGGT/DAv3->latent space->finetune->3D diffusion-based NVS model arxiv.org/abs/2603.22275

DualCoT-VLA: Visual-Linguistic Chain of Thought via Parallel Reasoning for Vision-Language-Action Models tl;dr: in title arxiv.org/abs/2603.22280

Rheos: Modelling Continuous Motion Dynamics in Hierarchical 3D Scene Graphs Iacopo Catalano, Francesco Verdoja, @jcivera.bsky.social, Jorge Peña-Queralta, Julio A. Placed tl;dr: continuous directional motion models->dynamics layer of a 3DSG arxiv.org/abs/2603.20239