Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

Bai et al., "Positional Encoding Field" Make your RoPE encoding 3D by including a z axis, then manipulate your image by simply manipulating your positional encoding in 3D --> novel view synthesis. Neat idea.

kwangmoo_yi's profile picture

Kwang Moo Yi

4,662 subscribers

46,846 Aufrufe • vor 7 Monaten •via X (Twitter)

BildungWissenschaft & Technologie
LIVE

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

✨ Live2D girl math ✨ If you add your Angle X group as a positional output to your Tracking Z group... And your Angle Z group as a positional output to your Tracking X group... You get: A wiggly tanuki! 🦝☕ [#Live2D WIP]
0:13

✨ Live2D girl math ✨ If you add your Angle X group as a positional output to your Tracking Z group... And your Angle Z group as a positional output to your Tracking X group... You get: A wiggly tanuki! 🦝☕ [#Live2D WIP]

Dawn ꒰🍃🍵꒱ DragonheartCafe | L2D opening March

28,154 Aufrufe • vor 4 Monaten

Discover the right 3D Geometric Foundation Model for your task—whether it’s stereo matching, multi-view depth estimation, video depth, pose estimation, semantic understanding, or novel view synthesis. Explore more insights in our #E3DBench #FoundationModel #3D #GaussianSplatting. Project Webpage:
0:14

Discover the right 3D Geometric Foundation Model for your task—whether it’s stereo matching, multi-view depth estimation, video depth, pose estimation, semantic understanding, or novel view synthesis. Explore more insights in our #E3DBench #FoundationModel #3D #GaussianSplatting. Project Webpage:

Zhiwen(Aaron) Fan

29,920 Aufrufe • vor 1 Jahr

[1/8] Drawing → 3D render with Gemini 2.0 image generation… by Dev Valladares + me 👇🧵Make your own at the link below
0:19

[1/8] Drawing → 3D render with Gemini 2.0 image generation… by Dev Valladares + me 👇🧵Make your own at the link below

Trudy Painter

71,766 Aufrufe • vor 1 Jahr

3D generation now down to 30s (was 5+ minutes before) using Z-Image T2I --> SAM-3D Image to 3D ⛏️
0:19

3D generation now down to 30s (was 5+ minutes before) using Z-Image T2I --> SAM-3D Image to 3D ⛏️

Blendi

127,707 Aufrufe • vor 6 Monaten

Meta's Segment Anything Model (SAM) can now run in your browser w/ WebGPU (+ fp16), meaning up to 8x faster image encoding (10s → 1.25s)! 🤯⚡️ Video is not sped up! Everything runs 100% locally thanks to 🤗 Transformers.js and onnxruntime-web! 🔗 Demo:
0:11

Meta's Segment Anything Model (SAM) can now run in your browser w/ WebGPU (+ fp16), meaning up to 8x faster image encoding (10s → 1.25s)! 🤯⚡️ Video is not sped up! Everything runs 100% locally thanks to 🤗 Transformers.js and onnxruntime-web! 🔗 Demo:

Xenova

120,352 Aufrufe • vor 2 Jahren

one 2d photo --> 3d gaussian splat quick test with Echo-2 by SpAItial -- these 3d scene generation models are getting better! already at a sufficient quality to serve as a virtual set / backdrop in your 3d tool of choice
0:17

one 2d photo --> 3d gaussian splat quick test with Echo-2 by SpAItial -- these 3d scene generation models are getting better! already at a sufficient quality to serve as a virtual set / backdrop in your 3d tool of choice

Bilawal Sidhu

23,511 Aufrufe • vor 1 Monat

🌐 N4E encoding exclusivity. decoding value. crafted for those who move different. built for crypto natives by natives. because true luxury isn't just appearance - it's a mindset. precision in every stitch. rebellion in every thread. not for everyone.
0:17

🌐 N4E encoding exclusivity. decoding value. crafted for those who move different. built for crypto natives by natives. because true luxury isn't just appearance - it's a mindset. precision in every stitch. rebellion in every thread. not for everyone.

not for everyone

164,175 Aufrufe • vor 1 Jahr

Have you 3D printed your free Walrus yet? 🦭 This ain’t just a file. It’s a symbol of the open manufacturing movement. A FREE 3D-printed drop—powered by your clicks, backed by the 3DOS DePIN army. Claim yours now →
0:10

Have you 3D printed your free Walrus yet? 🦭 This ain’t just a file. It’s a symbol of the open manufacturing movement. A FREE 3D-printed drop—powered by your clicks, backed by the 3DOS DePIN army. Claim yours now →

3DOS | Decentralized Manufacturing Network

127,774 Aufrufe • vor 1 Jahr

This is your sign to make a 3d animations Tiktok
0:13

This is your sign to make a 3d animations Tiktok

0xMarioNawfal

56,600 Aufrufe • vor 2 Monaten

Wonderland: Navigating 3D Scenes from a Single Image Contributions: • First, we introduce a representation for controllable 3D generation by leveraging the generative priors from camera-guided video diffusion models. Unlike image models, video diffusion models are trained on extensive video datasets. This enables them to capture comprehensive spatial relationships within scenes across multiple views and embed a form of "3D awareness" in their latent space, which allows us to maintain 3D consistency in novel view synthesis. • Second, to achieve controllable novel view generation, we empower video models with precise control over specified camera motions. We introduce a novel dual-branch conditioning mechanism that effectively incorporates desired diverse camera trajectories into the video diffusion model. This enables expansion of a single image into a multi-view consistent capture of a 3D scene with precise pose control. • Third, to achieve efficient 3D reconstruction, we directly transform video latents into 3DGS. We propose a novel latent-based large reconstruction model (LaLRM) that lifts video latents to 3D in a feed-forward manner. With this design, during inference, our model directly predicts 3DGS from a single input image, effectively aligning the generation and reconstruction tasks—and bridging image space and 3D space—through the video latent space. Compared with reconstructing scenes from images, the video latent space offers a 256× spatial-temporal reduction while retaining essential and consistent 3D structural details. Such a high degree of compression is crucial, as it allows the LaLRM to handle a wider range of 3D scenes within the reconstruction framework, with the same memory constraints.
0:12

Wonderland: Navigating 3D Scenes from a Single Image Contributions: • First, we introduce a representation for controllable 3D generation by leveraging the generative priors from camera-guided video diffusion models. Unlike image models, video diffusion models are trained on extensive video datasets. This enables them to capture comprehensive spatial relationships within scenes across multiple views and embed a form of "3D awareness" in their latent space, which allows us to maintain 3D consistency in novel view synthesis. • Second, to achieve controllable novel view generation, we empower video models with precise control over specified camera motions. We introduce a novel dual-branch conditioning mechanism that effectively incorporates desired diverse camera trajectories into the video diffusion model. This enables expansion of a single image into a multi-view consistent capture of a 3D scene with precise pose control. • Third, to achieve efficient 3D reconstruction, we directly transform video latents into 3DGS. We propose a novel latent-based large reconstruction model (LaLRM) that lifts video latents to 3D in a feed-forward manner. With this design, during inference, our model directly predicts 3DGS from a single input image, effectively aligning the generation and reconstruction tasks—and bridging image space and 3D space—through the video latent space. Compared with reconstructing scenes from images, the video latent space offers a 256× spatial-temporal reduction while retaining essential and consistent 3D structural details. Such a high degree of compression is crucial, as it allows the LaLRM to handle a wider range of 3D scenes within the reconstruction framework, with the same memory constraints.

MrNeRF

52,801 Aufrufe • vor 1 Jahr

🎶Time to start NeRFing audio. Introducing NVAS3D for novel acoustic synthesis in 3D scenes. Watch with AirPods. arxiv: code:
0:17

🎶Time to start NeRFing audio. Introducing NVAS3D for novel acoustic synthesis in 3D scenes. Watch with AirPods. arxiv: code:

Anurag Ranjan

24,702 Aufrufe • vor 2 Jahren

Combining the explicit control of 3D software with the creativity of generative AI models is a promising yet underrated workflow. Build your 3D scenes procedurally by describing them in natural language, then take them all the way with your image & video models of choice. Tools like intangible are built around such a workflow so you don't need to duct-tape apps together. Pretty cool!
0:15

Combining the explicit control of 3D software with the creativity of generative AI models is a promising yet underrated workflow. Build your 3D scenes procedurally by describing them in natural language, then take them all the way with your image & video models of choice. Tools like intangible are built around such a workflow so you don't need to duct-tape apps together. Pretty cool!

Bilawal Sidhu

37,544 Aufrufe • vor 11 Monaten

The way you get 3D shoulders is by developing your side and rear delts, not your front delts!
0:16

The way you get 3D shoulders is by developing your side and rear delts, not your front delts!

Up Workout

29,818 Aufrufe • vor 2 Monaten

Introduction to Positional Play | Concept 1⃣1⃣ - Through Circulation and Penetration, the Team Travels up the Field Together. Session Idea - Animations ⬇️
0:17

Introduction to Positional Play | Concept 1⃣1⃣ - Through Circulation and Penetration, the Team Travels up the Field Together. Session Idea - Animations ⬇️

Tom Skeath

19,862 Aufrufe • vor 1 Jahr

> become educated women by dad money > make your mom divorce you dad > make your mom marry another man, replacing your dad. OVARI DA
0:16

> become educated women by dad money > make your mom divorce you dad > make your mom marry another man, replacing your dad. OVARI DA

jeeta posting L (Old man era)

104,740 Aufrufe • vor 3 Monaten

Check out MultiDiff #CVPR2024! From a single RGB image, MultiDiff enables scene-level novel view synthesis with free camera control. Great work by @normanisation Katja Schwarz Barbara Roessle, L Porzi, S Rota Bulò, P Kontschieder
0:15

Check out MultiDiff #CVPR2024! From a single RGB image, MultiDiff enables scene-level novel view synthesis with free camera control. Great work by @normanisation Katja Schwarz Barbara Roessle, L Porzi, S Rota Bulò, P Kontschieder

Matthias Niessner

39,053 Aufrufe • vor 1 Jahr

Fun & likely unintended use case for #NanoBananaPro: image→3D point cloud Prompt: This is a rendering of a 3D model of a person. Make an accurate image of the corresponding normal map. Then I run Poisson-based normals→3D Pretty amazing for a general-purpose 2D image model 🤩
0:15

Fun & likely unintended use case for #NanoBananaPro: image→3D point cloud Prompt: This is a rendering of a 3D model of a person. Make an accurate image of the corresponding normal map. Then I run Poisson-based normals→3D Pretty amazing for a general-purpose 2D image model 🤩

Keenan Crane

71,412 Aufrufe • vor 6 Monaten

Imagine: Parallel Neural Networks with Shaders 🤔 (GLSL instead of CUDA) Parallelization by encoding the weights and activations as Buffer Objects or Textures, using fs to perform matmul and activation functions in parallel. Each fragm shader instance parproc 1..n neuron.
0:10

Imagine: Parallel Neural Networks with Shaders 🤔 (GLSL instead of CUDA) Parallelization by encoding the weights and activations as Buffer Objects or Textures, using fs to perform matmul and activation functions in parallel. Each fragm shader instance parproc 1..n neuron.

wavefnx

86,634 Aufrufe • vor 1 Jahr

Create Beautiful 3D in Seconds on Vision Pro. Introducing a new way to create playful 3D models on your Apple Vision Pro, generated by voice or text. Available now in Beta. Link in beautiful things bio.
0:15

Create Beautiful 3D in Seconds on Vision Pro. Introducing a new way to create playful 3D models on your Apple Vision Pro, generated by voice or text. Available now in Beta. Link in beautiful things bio.

Matt Hoerl

17,250 Aufrufe • vor 1 Jahr