Manuel Dahnert

Congrats to @Normanisation for his successful PhD defense 🥳🎓 Norman's thesis about 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐯𝐞 𝐌𝐨𝐝𝐞𝐥𝐬 𝐨𝐧 𝟑𝐃 𝐑𝐞𝐩𝐫𝐞𝐬𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧𝐬 makes important contributions to the 3D vision community. For instance, DiffRF, a generative approach directly operating in 3D space, was among the first diffusion techniques for neural radiance fields. This led to many follow up works in this area and sparked interest across the computer vision community, establishing generative approaches as a corner stone in the 3D domain. Also after his PhD, Norman continues to work on the forefront in computer vision, such as his contributions to MapAnything, a universal feedforward approach for 3D reconstruction. Check out Norman's amazing work: normanm.de Congratulations Dr. Mueller - super proud!

Check out or workshop on Generate Scene Completion at ICCV'25. We have an incredible speaker lineup and most certainly the coolest website (credits to @ethanjohnweber and @cursor_ai). 📅Mon, Oct 20 (morning session) 🌐scenecomp.github.io

Meet MapAnything – a transformer that directly regresses factored metric 3D scene geometry (from images, calibration, poses, or depth) in an end-to-end way. No pipelines, no extra stages. Just 3D geometry & cameras, straight from any type of input, delivering new state-of-the-art results 🚀 One universal model enables SoTA for: 🔥 Mono Depth Estimation 🔥 Multi-View SfM 🔥 Multi-View Stereo 🔥 Depth Completion 🔥 Registration … and many more possibilities! – plus everything is metric 🎯 We release code for data processing, training, benchmarking & ablations – everything Apache 2.0! Details & Links 👇

📢MeshPad: Interactive Sketch-Conditioned Artist-Designed Mesh Generation and Editing📢 Users can interactively design 3D models just from a sketch-based interface - check out the demo :) We break down the design process into addition with an autoregressive generator and deletion operations enabled by a classifier. To speed-up predictions, we propose a mesh-specific speculator such that users get immediate within a few seconds. Project: derkleineli.github.io/meshpad/ Video: youtu.be/_T6UTGTMZ1E Great work by @hcxrli @ErkocZiya @craigleili @DSirigatti V. Rosov @angelaqdai

📢Animating the Uncaptured 📢 We animate 3D humanoid meshes using video diffusion priors given a text prompt. 🎥youtu.be/_YL1J_V3smI 🌍marcb.pro/atu Realistic motion generation for 3D characters - without motion capture! 🚀 Great work by @marcbenedi @angelaqdai

Tomorrow in our TUM AI - Lecture Series with none other than Robin Rombach (@robrombach), CEO @bfl_ml. He'll talk about "𝐅𝐋𝐔𝐗: Flow Matching for Content Creation at Scale". Live stream: youtube.com/live/nrKKLJXBS… 6pm GMT+1 / 9am PST (Mon Feb 17rd)

Germans living abroad: last chance to register for the upcoming election is *tomorrow* (2 Feb 2025): auswaertiges-amt.de/de/service/269…

📢 ScanNet++ v2 Benchmark Release! 🏆 Test your state-of-the-art models on: 🔹 Novel View Synthesis 📸➡️🖼️ 🔹 3D Semantic & Instance Segmentation 🤖🔍🕶️ Shoutout to @chandan__yes and @liuyuehcheng for their incredible work👏 🚀Check it out: kaldir.vc.in.tum.de/scannetpp/

My group is looking for motivated PhD students that want to work on the future of digital humans. Within the ERC project 'LeMo: Learning Digital Humans in Motion' there are two open positions: career.tu-darmstadt.de/HPv3.Jobs/TU-D… career.tu-darmstadt.de/HPv3.Jobs/TU-D…

Great way to expand one's graphics knowledge: Every day grab and read a random graphics paper from realtimerendering.com/random.html. Even if you struggle to get the specifics, read the abstract or skim over it to learn about its existence and get the gist of a graphics technique.

Excited to announce ScanNet++ v2!🎉 @chandan__yes and @liuyuehcheng have been working tirelessly to bring: 🔹1006 high-fidelity 3D scans 🔹+ DSLR & iPhone captures 🔹+ rich semantics Elevating 3D scene understanding to the next level!🚀 w/ @MattNiessner kaldir.vc.in.tum.de/scannetpp

📢DNF: Generating 4D animations with dictionary-based neural fields! @xinyi092298 presents a new dictionary-based neural field for unconditional 4D generation of deforming shapes -- generating motions with high-quality shape and temporal consistency. xzhang-t.github.io/project/DNF/

📢📢𝐆𝐀𝐅: 𝐆𝐚𝐮𝐬𝐬𝐢𝐚𝐧 𝐀𝐯𝐚𝐭𝐚𝐫 𝐑𝐞𝐜𝐨𝐧𝐬𝐭𝐫𝐮𝐜𝐭𝐢𝐨𝐧 𝐟𝐫𝐨𝐦 𝐌𝐨𝐧𝐨𝐜𝐮𝐥𝐚𝐫 𝐕𝐢𝐝𝐞𝐨𝐬 𝐯𝐢𝐚 𝐌𝐮𝐥𝐭𝐢-𝐯𝐢𝐞𝐰 𝐃𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧📢📢 We reconstruct animatable Gaussian head avatars from monocular videos captured by commodity devices such as smartphones. Key idea: distill reconstruction constraints from a multi-view head diffusion model to complete unobserved regions. tangjiapeng.github.io/projects/GAF/ youtu.be/QuIYTljvhyg Great work by @jiapeng_tang @davidedavoli @TobiasKirschst1 @liamschoneveld

📢MeshArt: Generating Articulated Meshes with Structure-guided Transformers @DaoyiGao generates articulated meshes with a hierarchical transformer, modeling articulation-aware structures that guide mesh synthesis. w/ @yawarnihal @craigleili Project: daoyig.github.io/Mesh_Art/

Our work Meta 3D AssetGen on 3D shape generation will be presented at #NeurIPS2024, on Thursday afternoon session (12 Dec 4:30 p.m. PST — 7:30 p.m. PST) in East Exhibit Hall A-C, Poster #4609! I won't be attending, but Prof. Andrea Vedaldi will be there. Come say hi!

Learnings: - Formulate single image scene reconstruction as conditional diffusion processes - Learn a generative scene prior by denoising all objects simultaneously leads to more coherent arrangements - Probabilistic diffusion formulation is well suited for ambiguous cases in single image reconstruction task - Exploit the intermediate shape representation for efficient surface alignment loss for joint training - Represent objects using disentangled shape encoding results in in high-quality shapes Great collab with @angelaqdai, @Normanisation and @MattNiessner

Super happy to present our #NeurIPS paper 𝐂𝐨𝐡𝐞𝐫𝐞𝐧𝐭 𝟑𝐃 𝐒𝐜𝐞𝐧𝐞 𝐃𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧 𝐅𝐫𝐨𝐦 𝐚 𝐒𝐢𝐧𝐠𝐥𝐞 𝐑𝐆𝐁 𝐈𝐦𝐚𝐠𝐞 in Vancouver. Come to our poster #2804 on Wednesday 11am - 2pm in East Exhibit Hall A-C and say hi if you want to learn more about 3D Scene Diffusion! See you tomorrow! Project Page: manuel-dahnert.com/research/scene… Poster: neurips.cc/virtual/2024/p…

📢SceneFactor: Generating & editing 3D indoor scenes from text! @ABokhovkin presents a factored latent diffusion for controllable, large-scale scene synthesis -- decomposed into high-level semantic generation + geometric refinement w/ @QTDSMQ, @shubhtuls alexeybokhovkin.github.io/scenefactor/

We present Mesh2NeRF at #ECCV this afternoon (Wed Oct 2, 16:30 - 18:30). Poster Session: 4 Poster Board ID: 290 Come by and let’s chat about it!

(1/2) How to accelerate the reconstruction of 3D Gaussian Splatting? 3DGS-LM replaces the commonly used ADAM optimizer with a tailored Levenberg-Marquardt (LM). => We are 𝟑𝟎% 𝐟𝐚𝐬𝐭𝐞𝐫 𝐭𝐡𝐚𝐧 𝟑𝐃𝐆𝐒 for the same quality. lukashoel.github.io/3DGS-LM/ youtu.be/tDiGuGMssg8