Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

DistScene- turns a single image into an editable 3D scene. Uses sparse voxels, sparse DiT flow-matching models + LoRA. Based on TRELLIS.2 object-centric refinement restores details

Wildminder

12,376 subscribers

48,029 views • 21 days ago •via X (Twitter)

Arts Education Science & Technology

Anya Rossi• Live Now

Private livecam show

5 Comments

Thusatharan21 days ago

if the objects come out as separate meshes, blender kitbashing gets way faster

tvitzer tvitzerovich21 days ago

coming soon

燎原21 days ago

超级棒，我想他小到不需要独立显卡

Andrii Zadaianchuk 🇺🇦21 days ago

Nice paper! Is it possible to expend for robotics, when more inputs about the scene are available like depth or multi-view? In RecGen we showed that it leads to much more accurate object placements in the scene, needed for real-to-sim

Soroosh21 days ago

does it hold up on messy real photos or only clean ones? that's where these fall apart for me

Related Videos

GenRecon turns a handful of room photos into professional-grade 3D models. Nice for games, VR, and architectural design. - PBR meshes from sparse RGB images - Trellis.2+DINOv3 - partitions large scenes into overlapping chunks for global consistency

GenRecon turns a handful of room photos into professional-grade 3D models. Nice for games, VR, and architectural design. - PBR meshes from sparse RGB images - Trellis.2+DINOv3 - partitions large scenes into overlapping chunks for global consistency

Wildminder

13,654 views • 5 months ago

You can now train a 3D LoRA, exclusively on fal! TRELLIS.2 LoRA Trainer turns a folder of 3D assets into custom image-to-3D adapters, fine-tune shape and texture, then generate GLBs that match your style.

You can now train a 3D LoRA, exclusively on fal! TRELLIS.2 LoRA Trainer turns a folder of 3D assets into custom image-to-3D adapters, fine-tune shape and texture, then generate GLBs that match your style.

fal

50,927 views • 3 months ago

A fast way to turn your phone vids into editable 3D rooms: WorldSculpt. Converts cluttered multi-view video into editable meshes - extracts hundreds of separate object meshes under 75% occlusion - uses Pixal3D + DINOv3 via LoRA - no scene-level training.

A fast way to turn your phone vids into editable 3D rooms: WorldSculpt. Converts cluttered multi-view video into editable meshes - extracts hundreds of separate object meshes under 75% occlusion - uses Pixal3D + DINOv3 via LoRA - no scene-level training.

Wildminder

37,530 views • 1 month ago

Direct3D-S2 Gigascale 3D Generation Made Easy with Spatial Sparse Attention high resolution 3D generation from image

Direct3D-S2 Gigascale 3D Generation Made Easy with Spatial Sparse Attention high resolution 3D generation from image

AK

59,290 views • 1 year ago

MoVerse turns an image into a navigable 3D world. Runs at 8 FPS on a single RTX 4090.

MoVerse turns an image into a navigable 3D world. Runs at 8 FPS on a single RTX 4090.

AI Search

18,995 views • 4 months ago

3D scene reconstructions by NVIDIA. ArtiFixer - repairs artifacts and extends sparse views via Wan 2.1. - high-fidelity inpainting in occluded regions - gens hundreds of consistent frames in a single pass - 3D Gaussian Splatting for navigable scene reconstruction Makes the 3D environment look photorealistic and fully navigable for VR/AR. It basically turns a broken 3D model into a polished, professional scene.

3D scene reconstructions by NVIDIA. ArtiFixer - repairs artifacts and extends sparse views via Wan 2.1. - high-fidelity inpainting in occluded regions - gens hundreds of consistent frames in a single pass - 3D Gaussian Splatting for navigable scene reconstruction Makes the 3D environment look photorealistic and fully navigable for VR/AR. It basically turns a broken 3D model into a polished, professional scene.

Wildminder

16,086 views • 4 months ago

🚨 A single image → a fully editable UE5 world. Introducing Lumera: a turnkey pipeline that reconstructs an engine-native 3D scene directly from ONE image — not just pixels, point clouds, or Gaussians. It gives you: → object-level 3D meshes → editable 3D layouts → engine-native lights → HDR environment lighting → a scene ready for Unreal Engine 5 In other words: Image → 3D world → UE5. We believe the future of image-to-3D is not “generate something that looks 3D.” It is generate the actual world project. 🌐 📄

🚨 A single image → a fully editable UE5 world. Introducing Lumera: a turnkey pipeline that reconstructs an engine-native 3D scene directly from ONE image — not just pixels, point clouds, or Gaussians. It gives you: → object-level 3D meshes → editable 3D layouts → engine-native lights → HDR environment lighting → a scene ready for Unreal Engine 5 In other words: Image → 3D world → UE5. We believe the future of image-to-3D is not “generate something that looks 3D.” It is generate the actual world project. 🌐 📄

Hao Zhao

52,630 views • 2 months ago

🥉 3rd Place: Custom Universe Jake Stevens & Mauricio Pereira built a real-time engine that turns a single phone photo into a fully editable, photorealistic 3D scene.

🥉 3rd Place: Custom Universe Jake Stevens & Mauricio Pereira built a real-time engine that turns a single phone photo into a fully editable, photorealistic 3D scene.

ClaudeDevs

56,243 views • 4 months ago

Code released: "Code for SC-GS: Sparse-Controlled Gaussian Splatting for Editable Dynamic Scenes"

Code released: "Code for SC-GS: Sparse-Controlled Gaussian Splatting for Editable Dynamic Scenes"

MrNeRF

11,951 views • 2 years ago

📢📢GenRecon: Bridging Generative Priors for Multi-View 3D Scene Reconstruction📢📢 Reconstructing high-fidelity 3D scenes from sparse RGB input is hard. It needs a strong 3D prior! We reformulate multi-view scene reconstruction as conditional 3D generation over overlapping spatial chunks, lifting posed image features into a generative shape prior via 3D conditioning. As an example prior, we build on Trellis2, and train it such that its reconstruction is pixel aligned and matches from all views. GenRecon achieves unprecedented reconstruction quality from any sparse RGB input sequence, even from a phone capture. The reconstruction also includes PBR materials which facilitates relighting and virtual object insertion. Amazing work by Katharina Schmid, Nicolas von Lützow, Jozef, Angela Dai

📢📢GenRecon: Bridging Generative Priors for Multi-View 3D Scene Reconstruction📢📢 Reconstructing high-fidelity 3D scenes from sparse RGB input is hard. It needs a strong 3D prior! We reformulate multi-view scene reconstruction as conditional 3D generation over overlapping spatial chunks, lifting posed image features into a generative shape prior via 3D conditioning. As an example prior, we build on Trellis2, and train it such that its reconstruction is pixel aligned and matches from all views. GenRecon achieves unprecedented reconstruction quality from any sparse RGB input sequence, even from a phone capture. The reconstruction also includes PBR materials which facilitates relighting and virtual object insertion. Amazing work by Katharina Schmid, Nicolas von Lützow, Jozef, Angela Dai

Matthias Niessner

18,447 views • 5 months ago

🚨 A SINGLE IMAGE can now become an entire 3D WORLD. I’m talking about Hyper3D WorldGen. It follows a structured method where EVERY OBJECT is an individual editable 3D asset. It also reconstructs: → Spatial positions & relationships → Object arrangements → Collisions → Mass → Friction etc... On the other hand, 3D Designers take much time to do the same things Free Trail Link Below👇

🚨 A SINGLE IMAGE can now become an entire 3D WORLD. I’m talking about Hyper3D WorldGen. It follows a structured method where EVERY OBJECT is an individual editable 3D asset. It also reconstructs: → Spatial positions & relationships → Object arrangements → Collisions → Mass → Friction etc... On the other hand, 3D Designers take much time to do the same things Free Trail Link Below👇

SARAH

87,362 views • 1 month ago

Microsoft just dropped TRELLIS.2 This AI turns image into photoreal-ish 3D with full PBR Now with higher definition and hyper realism Code + live demo below 👇

Microsoft just dropped TRELLIS.2 This AI turns image into photoreal-ish 3D with full PBR Now with higher definition and hyper realism Code + live demo below 👇

Min Choi

103,328 views • 10 months ago

David Ambarzumjan is a Berlin-based artist whose Brushstrokes in Time series uses a single sweeping stroke to shift one scene into another, merging different moments and realities into a single image

David Ambarzumjan is a Berlin-based artist whose Brushstrokes in Time series uses a single sweeping stroke to shift one scene into another, merging different moments and realities into a single image

Science girl

222,497 views • 6 months ago

image → 3D generation. a prompt is gently refined with Nemotron → FLUX paints a reference image → TRELLIS reconstructs a textured 3D mesh, running locally on an NVIDIA AI DGX Spark. quiet, local, always on.

image → 3D generation. a prompt is gently refined with Nemotron → FLUX paints a reference image → TRELLIS reconstructs a textured 3D mesh, running locally on an NVIDIA AI DGX Spark. quiet, local, always on.

three.ws

34,487 views • 4 months ago

🚀Turn Single Image into 3D Human🚀 #GeneMAN is a generalizable single-image 3D human reconstruction framework that turns in-the-wild images into high-quality 3D humans with ease 🔗Project: 📜Paper: 🧑‍💻Code:

🚀Turn Single Image into 3D Human🚀 #GeneMAN is a generalizable single-image 3D human reconstruction framework that turns in-the-wild images into high-quality 3D humans with ease 🔗Project: 📜Paper: 🧑‍💻Code:

Ziwei Liu

26,953 views • 1 year ago

Efficient LoFTR was just integrated into Hugging Face Transformers! It improves upon LoFTR, a detector-free image matcher, by being 2.5x faster. It can even surpass the SOTA efficient sparse matching pipeline SuperPoint + LightGlue. Now available in a few lines of code!

Efficient LoFTR was just integrated into Hugging Face Transformers! It improves upon LoFTR, a detector-free image matcher, by being 2.5x faster. It can even surpass the SOTA efficient sparse matching pipeline SuperPoint + LightGlue. Now available in a few lines of code!

Niels Rogge

32,294 views • 1 year ago

ByteDance and Carnegie Mellon researchers just announced PartCrafter This AI turns a single photo into fully editable 3D parts in seconds 10 wild examples:

ByteDance and Carnegie Mellon researchers just announced PartCrafter This AI turns a single photo into fully editable 3D parts in seconds 10 wild examples:

Min Choi

963,037 views • 1 year ago

RIP to everyone who spent years learning 3D in Blender 💀 Hi3D just released an open-source model that turns any object from a photo into a full 3D model. It's trained on 2 million 3D assets, and in their technical report 90% of its models pass 3D-print slicing. Just generate the base from a photo, refine the details with Opus 5.5 and you're done. The barrier to 3D just disappeared, you can find more details in the article below:

RIP to everyone who spent years learning 3D in Blender 💀 Hi3D just released an open-source model that turns any object from a photo into a full 3D model. It's trained on 2 million 3D assets, and in their technical report 90% of its models pass 3D-print slicing. Just generate the base from a photo, refine the details with Opus 5.5 and you're done. The barrier to 3D just disappeared, you can find more details in the article below:

Anatoli Kopadze

212,593 views • 1 day ago

Making 3D assets just got stupidly easy. Tripo turns a sentence or a single image into a full 3d model, ready for blender, games, printing, whatever. Idea → image → 3D asset. Full workflow below 🧵

Making 3D assets just got stupidly easy. Tripo turns a sentence or a single image into a full 3d model, ready for blender, games, printing, whatever. Idea → image → 3D asset. Full workflow below 🧵

Farhan

44,468 views • 4 months ago