VisualitoXr

[VisualitoXr @VisualitoXr]

Try it here👉 https://t.co/7DHBui3rWk

[VisualitoXr @VisualitoXr]

🚀 SAM3 + SAM3D → 4DGS

Turning video into 4D Gaussian Splatting using Meta’s newest Segment Anything 3 + Segment Anything 3D models.

#SAM3 #SAM3D #4DGS #PlayCanvas #AIatMeta

[VisualitoXr @VisualitoXr]

Thermal 4D Gaussian Splatting 🌡️

📱 Capture: iPhone 16 Pro + Topdon Thermal Camera
⚙️ Processing: LichtFeld Studio
🎨 Visualization: SuperSplat

Interactive model here: 👇
https://t.co/1x4XZ7A38O

#LichtFeldStudio #PlayCanvas #GaussianSplatting

[VisualitoXr @VisualitoXr]

Thermal 3D Scanning 🌡️🏠

📱 Capture: iPhone 16 Pro + Thermal Camera
⚙️ Processing: LichtFeld Studio
📂 Output: 3D Gaussian Splats
🎨 Visualization: SuperSplat

Interactive model here: 👇
https://t.co/3r4dFaxqBY

#GaussianSplatting #LichtFeldStudio #PlayCanvas

[VisualitoXr @VisualitoXr]

Mobile LiDAR-SLAM Gaussian Splatting with Thermal Imaging

Captured on iPhone 16 Pro with a thermal camera, trained in LichtFeld Studio, and visualized in SuperSplat.

#PlayCanvas #LichtFeldStudio #GaussianSplatting

[VisualitoXr @VisualitoXr]

@migero_usausagi Yes, this is a limitation of the DA-2 algorithm: https://t.co/VDiTf8B1kh

Pano SHARP could potentially deliver better results; the published results look promising:
https://t.co/0rQFH3prIp

[Daniel Skaale @DSkaale]

🚀 Just created something completely new!

I Rewrote Apple's Sharp to generate 3D Gaussian Splats from a single 360° equirectangular image made from FLUX 2 (Flex)
One equirectangular image → Full spherical 3DGS world!

✨ New web interface - drag & drop your pano, select quality, download PLY ⚡ 2-3 sec on a older laptop

Quality tiers:
▫️ Low: 500K splats
▫️ Medium: 2M splats
▫️ High: 4.7M splats
▫️ Ultra: 8.4M splats

The pipeline:
360° depth estimation from equirectangular input
Spherical projection maps pixels → 3D points
Each point becomes a Gaussian with depth-scaled radius
- Export to standard PLY for any 3DGS viewer (Unity)

No multi-view capture. No COLMAP. No NeRF training.
Just upload a pano, pick quality, get your splats.

#3DGS #GaussianSplatting #ComputerVision #AI #madewithunity #Splats

Sharp Monocular View Synthesis in Less Than a Second https://t.co/eVLtXUlqQw

[VisualitoXr @VisualitoXr]

360° Video → 4DGS Panorama Splats

Insta360 X5 → equirectangular → DA-2 (per frame) → Gaussian splats → SOG (~3 MB/frame) → 4DGS panorama playback in SuperSplat (custom code)

Try it here👉 https://t.co/khNpczDRH9

#PlayCanvas #4DGS #360Video

[VisualitoXr @VisualitoXr]

@harry_bardak As expected for per-frame monocular depth estimation, near-field geometry is relatively stable, whereas far-field regions exhibit temporal drift and reduced temporal consistency across frames (see video)

[VisualitoXr @VisualitoXr]

Sharp Monocular View Synthesis in Less Than a Second

Monocular Video → 4DGS
Per-frame Gaussian splats generated from monocular video, visualized in the SuperSplat Editor.
SHARP project page: https://t.co/Ny3FNLMlUB

#AppleML #PlayCanvas #SHARP #GaussianSplatting

[VisualitoXr @VisualitoXr]

@JohnOfProgress Yes — the “animation” is simply a sequence of PLY files.
If you name them 00000.ply, 00001.ply, 00002.ply, … you can just drag & drop the entire sequence into SuperSplat.
SuperSplat interprets them as consecutive frames and plays them as a 4DGS animation.

[VisualitoXr @VisualitoXr]

From Mesh Splatting to a Textured Mesh ✨

Input images → triangle soup → Delaunay mesh reconstruction → simplification, smoothing, and texturing.

Custom Delaunay mesh reconstruction implementation inspired by the Mesh Splatting paper.

#MeshSplatting

[VisualitoXr @VisualitoXr]

Radiance Meshes: Quick Test 🔆🕸️

Watch the video and inspect the trained model in the browser with the online WebGPU Tetrahedral Viewer: 👉 https://t.co/4MSLppYWHL

Radiance Meshes Project Page: 🔗 https://t.co/wz7zA2Fiyc

#RadianceMeshes

[VisualitoXr @VisualitoXr]

360° PANORAMA With Nano Banana 🍌

Watch the video, generate your own panorama & try the demo: 👉 https://t.co/YJUNtt7YIM

⌨️ Try this Prompt:

"Create full 360 panorama of [YOUR PLACE AND TIME] with corresponding depth image"

#WebXR #VR #NanoBanana #AI

[VisualitoXr @VisualitoXr]

@taigajim 4DGS SAM3 & SAM3D pipeline:

1. Input video
2. SAM3 Python script (video + prompt) → saves each segment as an image for all video frames
3. SAM3D Python script → generates Gaussian Splatting frame-by-frame from those segments
4. Load multiple GS outputs in SuperSplat

[VisualitoXr @VisualitoXr]

@willeastcott I tried your SuperSplat for displaying and rendering 4D Gaussian Splatting — it works great!

https://t.co/sEmDW9bPvU

[VisualitoXr @VisualitoXr]

🚀 SAM3 + SAM3D → 4DGS

Turning video into 4D Gaussian Splatting using Meta’s newest Segment Anything 3 + Segment Anything 3D models.

#SAM3 #SAM3D #4DGS #PlayCanvas #AIatMeta

[VisualitoXr @VisualitoXr]

@AIatMeta SAM 3D is so cool 😎🧠✨ Great work!

https://t.co/toRFoWT0my

[VisualitoXr @VisualitoXr]

Watch this cool Meta SAM 3D demo — Photo → 3D Gaussian Splatting!
https://t.co/raOPpObupQ

#SAM3D #Meta

[VisualitoXr @VisualitoXr]

Watch this cool Meta SAM 3D demo — Photo → 3D Gaussian Splatting!
https://t.co/raOPpObupQ

#SAM3D #Meta

[VisualitoXr @VisualitoXr]

Experiments with Depth Anything 3 🚀 — experiments with depth estimation, camera position, point clouds and Gaussian splatting

#DepthAnything3