Загрузка видео...

Не удалось загрузить видео

На главную

My 3DGS tool [Extract Sharpest Frame] will support to generate masks by free text with SAM3 in next version. Now final testing. The flexible masking improves the result of SfM and 3DGS.😀 #3DGS #SAM3 #Meta

kotohibi_3d's profile picture

kotohibi

2,872 subscribers

21,541 просмотров • 3 месяцев назад •via X (Twitter)

#3DGS#SAM3#Meta
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

With DJI AVATA360 and 3DGS, let's be a bird😀. (4K) Tools : * AVATA360 : 10min fly, DLog * my 3DGS tools : frame extraction, SAM3 masking, cubemap conversion (now on sale) * Metashape std : SfM * LFS : 3DGS train with MRNF (no cleaning) * Postshot : Camera work
1:17

With DJI AVATA360 and 3DGS, let's be a bird😀. (4K) Tools : * AVATA360 : 10min fly, DLog * my 3DGS tools : frame extraction, SAM3 masking, cubemap conversion (now on sale) * Metashape std : SfM * LFS : 3DGS train with MRNF (no cleaning) * Postshot : Camera work

kotohibi

15,621 просмотров • 2 месяцев назад

I re-processed this 3DGS with dense splats and seam mask. (See in 4K) * Seam mask is to reduce errors around stitch line of dual fisheye for SfM and 3DGS train.(See comment) ---Steps--- 1. DJI OSMO360 10min walk 2. Frame extraction and SAM3 + seam masking : 774 images by my 3DGS tools 3. Spherical SfM by Metashape std with the mask 5. Cubemap conversion with the seam mask by my 3DGS tools 6. 3DGS train by LichtFeld Studio : 15M, SH2, MRNF, PPISP, RTX4090 10hours 7. Cam work by LichtFeld Studio ---------- #LFS #OSMO360
1:00

I re-processed this 3DGS with dense splats and seam mask. (See in 4K) * Seam mask is to reduce errors around stitch line of dual fisheye for SfM and 3DGS train.(See comment) ---Steps--- 1. DJI OSMO360 10min walk 2. Frame extraction and SAM3 + seam masking : 774 images by my 3DGS tools 3. Spherical SfM by Metashape std with the mask 5. Cubemap conversion with the seam mask by my 3DGS tools 6. 3DGS train by LichtFeld Studio : 15M, SH2, MRNF, PPISP, RTX4090 10hours 7. Cam work by LichtFeld Studio ---------- #LFS #OSMO360

kotohibi

12,874 просмотров • 1 месяц назад

I'm again reprocessing an old OSMO360 data with the latest my workflow (Dual SAM3 + Seam mask). See in 4K. --Steps-- 1. OSMO 360 10min walk ("RockSteady off") 2. Frame extraction and masking in my 3DGS tool : 613 images 3. SfM in Metashape std. with the masks -------- The sparse point cloud looks better than previous my workflow.😀
0:50

I'm again reprocessing an old OSMO360 data with the latest my workflow (Dual SAM3 + Seam mask). See in 4K. --Steps-- 1. OSMO 360 10min walk ("RockSteady off") 2. Frame extraction and masking in my 3DGS tool : 613 images 3. SfM in Metashape std. with the masks -------- The sparse point cloud looks better than previous my workflow.😀

kotohibi

21,779 просмотров • 28 дней назад

An entire cape 3DGS by AVATA360 & AprilTag [workflow] 12min fly -> my tool (Frame extraction&SAM3 masking) -> Metashape std -> my tool(Cubemap&Size estimation) -> LFS (5M splats & no clean) I will check the accuracy of size later on. See my guide how to make this😀
0:47

An entire cape 3DGS by AVATA360 & AprilTag [workflow] 12min fly -> my tool (Frame extraction&SAM3 masking) -> Metashape std -> my tool(Cubemap&Size estimation) -> LFS (5M splats & no clean) I will check the accuracy of size later on. See my guide how to make this😀

kotohibi

22,471 просмотров • 2 месяцев назад

This is a result of spherical SfM by Metashape std. It is very stable, accurate, fast and even no failure. Best match with AVATA 360, I think😀 -------------- Location : Miho dam (See the comment) Video length : 14:34 flight by DJI AVATA360 Frame extraction and masking : 875 images by my 3DGS tools ------------- Move to 3DGS train.💪
0:45

This is a result of spherical SfM by Metashape std. It is very stable, accurate, fast and even no failure. Best match with AVATA 360, I think😀 -------------- Location : Miho dam (See the comment) Video length : 14:34 flight by DJI AVATA360 Frame extraction and masking : 875 images by my 3DGS tools ------------- Move to 3DGS train.💪

kotohibi

20,505 просмотров • 1 месяц назад

A 3DGS of a dam in Japan by DJI AVATA 360 (See in 4K) ---steps--- 1. Location : Miho dam in Tanzawa lake 2. Video length : 14:34 8K DLog-M 3. Frame extraction and masking : 875 images by my 3DGS tools 4. Spherical SfM by Metashape std with mask 5. Cubemap conversion and parallax-based reduction by my 3DGS tools 6. 3DGS train by LichtFeld Studio : 15M, SH2, MRNF, PPISP 7. Cam work by LichtFeld Studio ------------- #LFS #丹沢湖 #三保ダム #AVATA360
0:59

A 3DGS of a dam in Japan by DJI AVATA 360 (See in 4K) ---steps--- 1. Location : Miho dam in Tanzawa lake 2. Video length : 14:34 8K DLog-M 3. Frame extraction and masking : 875 images by my 3DGS tools 4. Spherical SfM by Metashape std with mask 5. Cubemap conversion and parallax-based reduction by my 3DGS tools 6. 3DGS train by LichtFeld Studio : 15M, SH2, MRNF, PPISP 7. Cam work by LichtFeld Studio ------------- #LFS #丹沢湖 #三保ダム #AVATA360

kotohibi

18,091 просмотров • 1 месяц назад

A test of real-time transfer of images captured on Quest to a PC, where they are converted into 3D using Apple's SHARP. Currently, the final 3DGS output is displayed on the PC. The next step is to bring the result back into Quest for in-headset viewing. #3DGS #metaquest #AR
1:34

A test of real-time transfer of images captured on Quest to a PC, where they are converted into 3D using Apple's SHARP. Currently, the final 3DGS output is displayed on the PC. The next step is to bring the result back into Quest for in-headset viewing. #3DGS #metaquest #AR

Takashi Yoshinaga

11,575 просмотров • 4 месяцев назад

One more DJI AVATA360 3DGS is here ! (See in 4K video) The data was taken by イナダユウキ,too (*Skilled drone artist*) I'm horner to collaboration with him😀 "AVATA360->Metashape(Tie point clean)->my tools->(LFS)LichtFeld Studio(MRNF)" All around view is possible thank to drone. And I didn't even do any cleanup thank to LFS MRNF. #AVATA360 #3DGS #LFS #DJI
0:38

One more DJI AVATA360 3DGS is here ! (See in 4K video) The data was taken by イナダユウキ,too (*Skilled drone artist*) I'm horner to collaboration with him😀 "AVATA360->Metashape(Tie point clean)->my tools->(LFS)LichtFeld Studio(MRNF)" All around view is possible thank to drone. And I didn't even do any cleanup thank to LFS MRNF. #AVATA360 #3DGS #LFS #DJI

kotohibi

46,240 просмотров • 4 месяцев назад

3DGS of DJI AVATA360 is here again(See in 4K video) Title : Mountain of stonecutting This data was taken by Taka🍥🍥 (*Skilled drone artist*) I'm honored to collaborate with him again😀. "AVATA360->Metashape ->my tools->(LFS) LichtFeld Studio (MRNF)" Using a 360-degree drone, you can capture the entire location and take a virtual tour😄. #AVATA360 #3DGS #LFS #DJI
1:21

3DGS of DJI AVATA360 is here again(See in 4K video) Title : Mountain of stonecutting This data was taken by Taka🍥🍥 (*Skilled drone artist*) I'm honored to collaborate with him again😀. "AVATA360->Metashape ->my tools->(LFS) LichtFeld Studio (MRNF)" Using a 360-degree drone, you can capture the entire location and take a virtual tour😄. #AVATA360 #3DGS #LFS #DJI

kotohibi

21,927 просмотров • 4 месяцев назад

3DGS of DJI AVATA360 is here again! (See in 4K videos) Title : A Ruined School This data was taken by Taka🍥🍥 (*Skilled drone artist*) I'm honored to collaborate with him😀 "AVATA360->Metashape (Tie point clean) ->my tools->(LFS) LichtFeld Studio (MRNF)" This scene involves frequent changes in exposure due to entering and exiting a building. Unfortunately, there is many floaters around the windows, but LFS output a pretty good quality even without cleaning because this is a so difficult scene for 3DGS. #AVATA360 #3DGS #LFS #DJI
1:54

3DGS of DJI AVATA360 is here again! (See in 4K videos) Title : A Ruined School This data was taken by Taka🍥🍥 (*Skilled drone artist*) I'm honored to collaborate with him😀 "AVATA360->Metashape (Tie point clean) ->my tools->(LFS) LichtFeld Studio (MRNF)" This scene involves frequent changes in exposure due to entering and exiting a building. Unfortunately, there is many floaters around the windows, but LFS output a pretty good quality even without cleaning because this is a so difficult scene for 3DGS. #AVATA360 #3DGS #LFS #DJI

kotohibi

22,773 просмотров • 4 месяцев назад

Painting with 3D Gaussian Splat Brushes Contributions: • Designing a set of interactive tools and brush parameters for artistic brush creation and control using 3DGS content. • Computing oriented 3D Gaussian splat brushes for stamp-based painting on 3D surfaces represented as meshes or 3DGS scenes. • Deforming the splats in a brush stamp to ensure a smooth appearance of the painted 3D stroke. • Producing seamless brush strokes despite overlapping brush stamps, using diffusion inpainting. • Efficient modeling and rendering of brush strokes to facilitate 3DGS painting in real time.
0:27

Painting with 3D Gaussian Splat Brushes Contributions: • Designing a set of interactive tools and brush parameters for artistic brush creation and control using 3DGS content. • Computing oriented 3D Gaussian splat brushes for stamp-based painting on 3D surfaces represented as meshes or 3DGS scenes. • Deforming the splats in a brush stamp to ensure a smooth appearance of the painted 3D stroke. • Producing seamless brush strokes despite overlapping brush stamps, using diffusion inpainting. • Efficient modeling and rendering of brush strokes to facilitate 3DGS painting in real time.

MrNeRF

23,010 просмотров • 1 год назад

A quick first pass at a new iteration of the realtime diffusion pipeline: physical interface → camera → realtime img2img → upscale → 3DGS through the new TripoSplat. Rough and held together with tape for now, but the whole loop runs end to end. Next: - better input and camera's pov, - splitting it across two machines to automate the hand-off, - and adding in the Tripo3D API for higher-res models... the 3D printer is just waiting for that. #diffusion #3DGS #TripoSplat
1:24

A quick first pass at a new iteration of the realtime diffusion pipeline: physical interface → camera → realtime img2img → upscale → 3DGS through the new TripoSplat. Rough and held together with tape for now, but the whole loop runs end to end. Next: - better input and camera's pov, - splitting it across two machines to automate the hand-off, - and adding in the Tripo3D API for higher-res models... the 3D printer is just waiting for that. #diffusion #3DGS #TripoSplat

Hugues Bruyère

15,526 просмотров • 2 месяцев назад

This is by far the best 3DGS scan of my kitchen. I love how the light reflects on my dining room table from the bay windows. I still struggle to be able to deal with the high dynamic range between the inside and outside though. The windows are overexposed which means you can't see anything outside well. Does anyone have a good solution to this? Shoot raw and globally adjust? I would love to share notes with anyone working on this challenge at #CVPR2026 #3DGS #ComputerVision
0:42

This is by far the best 3DGS scan of my kitchen. I love how the light reflects on my dining room table from the bay windows. I still struggle to be able to deal with the high dynamic range between the inside and outside though. The windows are overexposed which means you can't see anything outside well. Does anyone have a good solution to this? Shoot raw and globally adjust? I would love to share notes with anyone working on this challenge at #CVPR2026 #3DGS #ComputerVision

Jonathan Stephens

22,090 просмотров • 3 месяцев назад

This is the biggest announcement from Apple that nobody has talked about. Generating #GaussianSplatting with #AI using models like Trellis or Sam3D hasn’t made much sense...until now. Meshes are still far easier to edit, integrate cleanly into #3D workflows, and with recent retopology advances, you can get high-quality low-poly assets with detailed textures. That said, this is where the balance starts to shift toward 3DGS. Until recently, AI struggled to generate splats with proper view-dependent effects. That’s changed with LiTo, a new model from Apple that can generate 3DGS with realistic view-dependent rendering, and the results are impressive. They’re not ready for game development pipelines yet. But for VFX, e-commerce, configurators, architecture, and similar domains, the implications are significant. What would it take for you to seriously explore AI-generated 3DGS for non-gaming use cases?
0:32

This is the biggest announcement from Apple that nobody has talked about. Generating #GaussianSplatting with #AI using models like Trellis or Sam3D hasn’t made much sense...until now. Meshes are still far easier to edit, integrate cleanly into #3D workflows, and with recent retopology advances, you can get high-quality low-poly assets with detailed textures. That said, this is where the balance starts to shift toward 3DGS. Until recently, AI struggled to generate splats with proper view-dependent effects. That’s changed with LiTo, a new model from Apple that can generate 3DGS with realistic view-dependent rendering, and the results are impressive. They’re not ready for game development pipelines yet. But for VFX, e-commerce, configurators, architecture, and similar domains, the implications are significant. What would it take for you to seriously explore AI-generated 3DGS for non-gaming use cases?

Gabriele Romagnoli

20,817 просмотров • 4 месяцев назад

The most incredible 3DGS you will see today 🤯 A 20 million Gaussian scan by ダックビル＠STUDIO DUCKBILL LLC of a primeval moss forest in Japan. 🌳 The sheer scale of splats continues to grow, covering ever larger areas. And this experience can be delivered to any device with a web browser. 🔥
1:16

The most incredible 3DGS you will see today 🤯 A 20 million Gaussian scan by ダックビル＠STUDIO DUCKBILL LLC of a primeval moss forest in Japan. 🌳 The sheer scale of splats continues to grow, covering ever larger areas. And this experience can be delivered to any device with a web browser. 🔥

Will Eastcott

11,004 просмотров • 2 месяцев назад

The Ring (Montreal) — Dormant Memories series New alternate scene, and an iteration on the UX and flow. Starting from a volumetric capture of a real location in Montreal, reconstructed as 3D Gaussian Splats, I then used World Labs' Marble to generate alternate versions of that same space. Running in real time using spark. #GaussianSplatting #3DGS #pointcloud #sparkjs #threejs
0:59

The Ring (Montreal) — Dormant Memories series New alternate scene, and an iteration on the UX and flow. Starting from a volumetric capture of a real location in Montreal, reconstructed as 3D Gaussian Splats, I then used World Labs' Marble to generate alternate versions of that same space. Running in real time using spark. #GaussianSplatting #3DGS #pointcloud #sparkjs #threejs

Hugues Bruyère

34,463 просмотров • 5 месяцев назад

oh may.. 3D Gaussian Splat is so crazy now last time we need weeks to turn a room into a 3D model.. now with AI 3DGS, you do it in seconds… even zoom into the text and looks perfect give it like 6 months… you can turn a whole city into a full detailed 3D model
0:22

oh may.. 3D Gaussian Splat is so crazy now last time we need weeks to turn a room into a 3D model.. now with AI 3DGS, you do it in seconds… even zoom into the text and looks perfect give it like 6 months… you can turn a whole city into a full detailed 3D model

el.cine

143,676 просмотров • 9 месяцев назад

Fun weekend project: using AI at Meta Sam3 to segment objects inside a SpAItial AI scene. Pipeline: -> Generate a world on SpAItial AI app and export as .ply -> PlayCanvas viewer to render splat + a uniform spatial grid over splat centers -> Capture frames + camera pose, POST to a local SAM3 server -> Lift each 2D mask → 3D: frustum-cull splats via the grid, reproject with the pose -> Recolor / isolate the segmented Gaussians This is all running locally on my Mac. Static scene, segment once, reuse persisted mask. All open source (link in comment), you can clone and build on top of this 👇🏽
1:17

Fun weekend project: using AI at Meta Sam3 to segment objects inside a SpAItial AI scene. Pipeline: -> Generate a world on SpAItial AI app and export as .ply -> PlayCanvas viewer to render splat + a uniform spatial grid over splat centers -> Capture frames + camera pose, POST to a local SAM3 server -> Lift each 2D mask → 3D: frustum-cull splats via the grid, reproject with the pose -> Recolor / isolate the segmented Gaussians This is all running locally on my Mac. Static scene, segment once, reuse persisted mask. All open source (link in comment), you can clone and build on top of this 👇🏽

RO2⚡️

22,119 просмотров • 2 месяцев назад

Our new Gaussian point splatting renders huge 3DGS scenes in real time (e.g. 425 M⬇️). It samples pixel-sized points from the Gaussians with massive parallelism and splats them to the screen using 64-bit atomics. Grab ACM SIGGRAPH 2026 paper+code+video now:
0:24

Our new Gaussian point splatting renders huge 3DGS scenes in real time (e.g. 425 M⬇️). It samples pixel-sized points from the Gaussians with massive parallelism and splats them to the screen using 64-bit atomics. Grab ACM SIGGRAPH 2026 paper+code+video now:

Christoph Peters

22,818 просмотров • 3 месяцев назад

🙏Thanks I'm overwhelmed by all the response!😍 ✨ Quick test with Pano360Sharp Just playing around with the output... 4M splats Added my Unity re-lighting and that dreamy tilt-shift blur effect 🎬 There's something magical about turning a single 360° photo into a little 3D world you can fly through 2-3 seconds to generate! #3DGS #GaussianSplatting #ComputerVision #AI #madewithunity #Splats
0:24

🙏Thanks I'm overwhelmed by all the response!😍 ✨ Quick test with Pano360Sharp Just playing around with the output... 4M splats Added my Unity re-lighting and that dreamy tilt-shift blur effect 🎬 There's something magical about turning a single 360° photo into a little 3D world you can fly through 2-3 seconds to generate! #3DGS #GaussianSplatting #ComputerVision #AI #madewithunity #Splats

Daniel Skaale

25,097 просмотров • 8 месяцев назад