When I said we'd smash the gangs, I meant the gangs of old ladies holding cardboard signs.
10.08.2025 09:46 โ ๐ 722 ๐ 193 ๐ฌ 14 ๐ 7@violetilla.bsky.social
Loud Spaniard. Maths & Software fan ๐ฉ๐ปโ๐ป๐ฑ๐ป Dr in Computer Vision. Working on AI in media production ๐ฌ Mostly rants about AI, ML, FOSS, science, ethics, politics & public transport in the UK ๐ดโโ๏ธ๐ฅพ๐ She/her
When I said we'd smash the gangs, I meant the gangs of old ladies holding cardboard signs.
10.08.2025 09:46 โ ๐ 722 ๐ 193 ๐ฌ 14 ๐ 7๐ New SIGGRAPH 2025 paper ๐
โHow to compute bounding volumes for procedural Signed Distance Fields (SDFs)? This is not so trivial!
๐กWe propose a simple method called Sphere Carving. It extracts (convex) bounding volumes around SDFs, requires very few evaluations, and is GPU compatible.
Voy segundo (!!) en el torneo de @kikollan.llaneras.es www.metaculus.com/c/kiko/?page=4 ๐ฑ
10.08.2025 12:47 โ ๐ 1 ๐ 1 ๐ฌ 0 ๐ 0The UK age verification rollout, chaotic as it is, is a proving ground for platforms that are looking ahead to using it on a global scale. In the US, thereโs never been a better time to get educated and get loud about the dangers of this legislation. www.eff.org/deeplinks/2...
09.08.2025 04:03 โ ๐ 151 ๐ 56 ๐ฌ 0 ๐ 6"Ethiopiaโs Tigray war, which lasted from 2020 to 2022, is one of the most criminally under-covered conflicts in modern history"
โIt is a really horrific and extreme form of sexual violence, and one that deserves the worldโs attention.โ๐งต
MuGS: Multi-Baseline Generalizable Gaussian Splatting Reconstruction
Yaopeng Lou, Liao Shen, Tianqi Liu, Jiaqi Li, Zihao Huang, Huiqiang Sun, Zhiguo Cao
tl;dr: DAv2 helps MVS via projected depth and sampled depth fusion for fine probability volume
arxiv.org/abs/2508.04297
RLGS: Reinforcement Learning-Based Adaptive Hyperparameter Tuning for Gaussian Splatting
Zhan Li, Huangying Zhan, Changyang Li, Qingan Yan, Yi Xu
tl;dr: hyperparameter tuning in 3DGS->Markov Decision Process->two policies->learning-rate scaling+densification adjustment
arxiv.org/abs/2508.04078
Surf3R: Rapid Surface Reconstruction from Sparse RGB Views in Seconds
Haodong Zhu, Changbai Li, Yangyang Ren, Zichao Feng, Xuhui Liu, Hanlin Chen, Xiantong Zhen, Baochang Zhang
tl;dr: DUSt3R+multiple reference views
arxiv.org/abs/2508.04508
๐ฃ๐ฃYou're in luck today. The deadline for the full paper submission has been extended for another two weeks until the 22nd of August!
We hope you're enjoying your holidays and come back with energy for research and fun!
Uni3R: Unified 3D Reconstruction and Semantic Understanding via Generalizable Gaussian Splatting from Unposed Multi-View Images
Xiangyu Sun, Haoyi jiang, Liu Liu, Seungtae Nam, Gyeongjin Kang, Xinjie wang, Wei Sui, Zhizhong Su, Wenyu Liu, Xinggang Wang, Eunbyung Park
arxiv.org/abs/2508.03643
Let's empty the jails so we can fill them with octogenarians who care about genocide.
Because that's the real threat to society.
www.theguardian.com/society/2025...
No Pose at All: Self-Supervised Pose-Free 3D Gaussian Splatting from Sparse Views
Ranran Huang, Krystian Mikolajczyk
tl;dr: NoPoSplat+Pose Head+reprojection loss
No idea what the main difference compared to NoPoSplat, but ICCV 2025 Highlight
Maybe good performance
arxiv.org/abs/2508.01171
Dens3R: A Foundation Model for 3D Geometry Prediction
Xianze Fang et 10 al.
tl;dr: hi-res, symmetric arch, "predict everything", version of DUST3R. No IMC eval.
arxiv.org/abs/2507.16290
Don't use a free VPN though, they sell your data and it's not safe. Do some research to see what the paid options are. Lots of benefits from having a good one ๐๐ฝ
05.08.2025 08:01 โ ๐ 4 ๐ 0 ๐ฌ 1 ๐ 0The biggest update in 3D reconstruction world - VGGT (new weights) has been released for commercial usage as well.
Kudos to @jianyuanwang.bsky.social to make this happen!
github.com/facebookrese...
MonoFusion: Sparse-View 4D Reconstruction via Monocular Fusion
Zihan Wang, Jeff Tan, @tarashakhurana.bsky.social, Neehar Peri, Deva Ramanan
tl;dr: align monocular reconstructions (from MoGe) to global reference frame (from DUSt3R) by adjusting the scale and shift
arxiv.org/abs/2507.23782
The anxiety of academic publishing ๐
31.07.2025 18:00 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0Reconstructing 4D Spatial Intelligence: A Survey
Yukang Cao, Jiahao Lu, Zhisheng Huang, Zhuowei Shen, Chengfeng Zhao, Fangzhou Hong, Zhaoxi Chen, Xin Li, Wenping Wang, Yuan Liu, Ziwei Liu
tl;dr: in title
arxiv.org/abs/2507.21045
Gaussian Set Surface Reconstruction through Per-Gaussian Optimization
Zhentao Huang, Di Wu, Zhenbang He, Minglun Gong
tl;dr: Gaussian centers->latent surface; align dominant normals with surface normal
arxiv.org/abs/2507.18923
G2S-ICP SLAM: Geometry-aware Gaussian Splatting ICP SLAM
Gyuhyeon Pak, Hae Min Cho, Euntai Kim
tl;dr: surface-aligned 2D Gaussian disk->anisotropic covariance prior->generalized ICP
arxiv.org/abs/2507.18344
Unposed 3DGS Reconstruction with Probabilistic Procrustes Mapping
Chong Cheng, Zijian Wang, Sicheng Yu, Yu Hu, Nanjie Yao, Hao Wang
tl;dr: submap alignment->point cloud registration->robust Umeyama algorithm->global point cloud and camera trajectory
arxiv.org/abs/2507.18541
In which I get both furious and very worried about google's AI putting incendiary falsehoods about evicting council tenants to house asylum seekers at the top of the search results. This is dangerous stuff (albeit unintended, though that might be worse).
nearlylegal.co.uk/2025/07/arti...
Seguimos rompiendo los muros de silencio gracias a medios como @elsaltodiario.com
El Festival de Cine #FiSahara pide a #ChristopherNolan que detenga el rodaje de โLa Odiseaโ
#TheOdyssey
www.elsaltodiario.com/sahara-occid...
๐๐๐
25.07.2025 10:29 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0Diffuman4D: 4D Consistent Human View Synthesis from Sparse-View Videos with Spatio-Temporal Diffusion Models
Yudong Jin, Sida Peng, Xuan Wang, Tao Xie, Zhen Xu, Yifan Yang, Yujun Shen, Hujun Bao, Xiaowei Zhou
arxiv.org/abs/2507.13344
Tendrรญa que ser ilegal cerrar un sitio como el Cafรฉ Central
24.07.2025 10:24 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0Still time to submit your papers to the 22nd ACM SIGGRAPH European Conference on Visual Media Production 2025 ๐ฅ๐๐ค
#CVMP2025
All dates: www.cvmp-conference.org/2025/call-fo...
IMPORTANT ANNOUNCEMENT ๐ฃ ๐ฃ
Sligh CVMP 2025 date change due to a venue issue. The confernece is now happening one day earlier.
The new dates are 3-4 of December ๐
Apologies if this causes any disruption to your plans
We hope to see you all there at #BFISouthbank this December! ๐ฅ