🎉 Thrilled to share our CVPR 2025 Award Candidate & Oral paper:. 🔹 GlobustVP.Convex Relaxation for Robust Vanishing Point Estimation in Manhattan World. 🚀 A globally optimal & outlier-robust method for vanishing point (VP) estimation. 🧱 Global optimality.💥 Tolerates up to

Enhancing Novel View Synthesis from extremely sparse views with SfM-free 3D Gaussian Splatting Framework. Zongqi He, et al. tl;dr: dense stereo (no introduction in paper, maybe *3R?)->initialization; training views->B-spline interpolation->video diffusion model->virtual

What do you want to know?.

Fast globally optimal Truncated Least Squares point cloud registration with fixed rotation axis. Ivo Ivanov, Carsten Markgraf. tl;dr: TLS->formulation; BnB->solver; contractor->reduce search space of BnB.

GSFix3D: Diffusion-Guided Repair of Novel Views in Gaussian Splatting. Jiaxin Wei, @StefanLeuteneg1, Simon Schaefer. tl;dr: fuse mesh and 3DGS->rendered images->pretrained diffusion model+random mask augmentation->removes artifacts+inpainting+completion.

Online 3D Gaussian Splatting Modeling with Novel View Selection. Byeonggwon Lee, Junkyu Park, Khang Truong Giang, Soohwan Song. tl;dr: shapes and positional gradients of Gaussians.->uncertainty->NBV->3DGS.

Distilled-3DGS:Distilled 3D Gaussian Splatting. Lintao Xiang, Xinkai Chen, Jianhuang Lai, @GuangcongW. tl;dr: vanilla 3DGS+noise-augmented 3DGS+dropout-regularized 3DGS->multi teachers->voxel histogram-based structural loss->student.

LongSplat: Robust Unposed 3D Gaussian Splatting for Casual Long Videos. @linjohnss, Cheng Sun, @FuEnYang1, @CMHungSteven, Yen-Yu Lin, @YuLunLiu790407. tl;dr: MASt3R+octree from F2-NeRF+Scaffold-GS for initialization; MASt3R+PnP+RANSAC for pose estimation; visibility local

4DNeX: Feed-Forward 4D Generative Modeling Made Easy. @Frozen_Burning, @TianqiLiu664, Long Zhuo, @jiawei6_ren, @zengtao197, He Zhu, @hongfz16, @pldeqiushui, @liuziwei7. tl;dr: 4D dataset; feed-forward framework for generating 4D scene representations from a single image by

G-CUT3R: Guided 3D Reconstruction with Camera and Depth Prior Integration. Ramil Khafizov, Artem Komarichev, @rusrakhimov, @peter_wonka, @burnaevevgeny. tl;dr: camera parameters or/and depthmaps->CUT3R.

i2Nav-Robot: A Large-Scale Indoor-Outdoor Robot Dataset for Multi-Sensor Fusion Navigation and Mapping. Hailiang Tang, Tisheng Zhang, Liqiang Wang, Xin Ding, Man Yuan, Zhiyu Xiang, Jujin Chen, Yuhan Bian, Shuangyan Liu, Yuqing Wang, Guan Wang, Xiaoji Niu. tl;dr: in title

ViPE: Video Pose Engine for 3D Geometric Perception. @huangjh_hjh, @QunjieZhou, @HesamRabeti, Aleksandr Korovko, @HuanLing6, @xuanchi13, @TianchangS, @JunGao33210520, Dmitry Slepichev, @chenhsuanlin, @jiawei6_ren, Kevin Xie, @Joydeepb_robots, @lealtaixe, @FidlerSanja. tl;dr:

BridgeDepth: Bridging Monocular and Stereo Reasoning with Latent Alignment. Tongfan Guan, Jiaxin Guo, @DrChenWang, Yun-Hui Liu. tl;dr: cross-attentive alignment dynamically synchronizes monocular contextual cues with stereo hypothesis representations.

STream3R: Scalable Sequential 3D Reconstruction with Causal Transformer. @GROS17121524, @TheYihangLuo, @hongfz16, @ShangchenZhou, Honghua Chen, Zhaoyang Lyu, Shuai Yang, @doubledaibo, @ccloy, @XingangP. tl;dr: 3D reconstruction->decoder-only transformer->sequential registration

ReconVLA: Reconstructive Vision-Language-Action Model as Effective Robot Perceiver. Wenxuan Song, Ziyang Zhou, Han Zhao, Jiayi Chen, Pengxiang Ding, Haodong Yan, Yuxin Huang, Feilong Tang, Donglin Wang, Haoang Li. tl;dr: reconstruction of gaze regions prompts the VLA model to

Puppeteer: Rig and Animate Your 3D Models. @chaoyue_song, Xiu Li, Fan Yang, @zhongcongxu, Jiacheng Wei, Fayao Liu, Jiashi Feng, Guosheng Lin, Jianfeng Zhang. tl;dr: integrate automatic rigging and animation into a unified pipeline.

Event-driven Robust Fitting on Neuromorphic Hardware. Tam Ngoc-Bang Nguyen, @DzungDoan6, @cai_zhipeng, @tatjunchin. tl;dr: spiking neural network->robust fitting on real neuromorphic hardware.

GSFixer: Improving 3D Gaussian Splatting with Reference-Guided Video Diffusion Priors. Xingyilang Yin, Qi Zhang, et al. tl;dr: 2D semantic (from DINOv2)+3D geometric (from VGGT) tokens from input views->video diffusion model->artifact-prone novel views; reference trajectory

A Survey on 3D Gaussian Splatting Applications: Segmentation, Editing, and Generation. Shuting He, Peilin Ji, Yitong Yang, Changshuo Wang, Jiayi Ji, Yinglin Wang, Henghui Ding. tl;dr: in title.

RayletDF: Raylet Distance Fields for Generalizable 3D Surface Reconstruction from Point Clouds or Gaussians. Shenxing Wei, Jinxi Li, Yafei Yang, Siyuan Zhou, @boyang_vLAR. tl;dr: raylet distance field->signed distance between the surface hit point and the raylet starting point

GaussianUpdate: Continual 3D Gaussian Splatting Update for Changing Environments. Lin Zeng, Boming Zhao, Jiarui Hu, Xujie Shen, Ziqiang Dang, Hujun Bao, @ZhaopengCui. tl;dr: hash encoding attribute updating+3D change detection+COLMAP-based point addition->changes;

DIP-GS: Deep Image Prior For Gaussian Splatting Sparse View Recovery. @rajaei_khatib, @RGiryes. tl;dr: re-purpose the inherent structure-inducing capabilities of DIP’s CNN to generate a structured 2D grid of 3D Gaussian parameters.