Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

VideoMaMa from KAIST: A mask-to-matte model, converts coarse video segmentation into pixel-accurate alpha mattes; looks quite good. Based on SVD+DINOv3.

wildmindai's profile picture

Wildminder

5,222 subscribers

19,652 Aufrufe • vor 7 Monaten •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

Create compositing mattes in seconds with Aleph 2.0 Mattes let you isolate a subject from its background so you can composite, color or apply effects to just one part of your shot. 1. Upload your video 2. Prompt for a white silhouette on black 3. Review your preview 4. Hit generate 5. Set your matte clip as a luma matte in your editor Clean mattes, ready for compositing. No rotoscoping required.
0:56

Create compositing mattes in seconds with Aleph 2.0 Mattes let you isolate a subject from its background so you can composite, color or apply effects to just one part of your shot. 1. Upload your video 2. Prompt for a white silhouette on black 3. Review your preview 4. Hit generate 5. Set your matte clip as a luma matte in your editor Clean mattes, ready for compositing. No rotoscoping required.

Runway

41,264 Aufrufe • vor 2 Monaten

Tracking Anything with Decoupled Video Segmentation paper page: Training data for video segmentation are expensive to annotate. This impedes extensions of end-to-end algorithms to new video segmentation tasks, especially in large-vocabulary settings. To 'track anything' without training on video data for every individual task, we develop a decoupled video segmentation approach (DEVA), composed of task-specific image-level segmentation and class/task-agnostic bi-directional temporal propagation. Due to this design, we only need an image-level model for the target task (which is cheaper to train) and a universal temporal propagation model which is trained once and generalizes across tasks. To effectively combine these two modules, we use bi-directional propagation for (semi-)online fusion of segmentation hypotheses from different frames to generate a coherent segmentation. We show that this decoupled formulation compares favorably to end-to-end approaches in several data-scarce tasks including large-vocabulary video panoptic segmentation, open-world video segmentation, referring video segmentation, and unsupervised video object segmentation.
0:32

Tracking Anything with Decoupled Video Segmentation paper page: Training data for video segmentation are expensive to annotate. This impedes extensions of end-to-end algorithms to new video segmentation tasks, especially in large-vocabulary settings. To 'track anything' without training on video data for every individual task, we develop a decoupled video segmentation approach (DEVA), composed of task-specific image-level segmentation and class/task-agnostic bi-directional temporal propagation. Due to this design, we only need an image-level model for the target task (which is cheaper to train) and a universal temporal propagation model which is trained once and generalizes across tasks. To effectively combine these two modules, we use bi-directional propagation for (semi-)online fusion of segmentation hypotheses from different frames to generate a coherent segmentation. We show that this decoupled formulation compares favorably to end-to-end approaches in several data-scarce tasks including large-vocabulary video panoptic segmentation, open-world video segmentation, referring video segmentation, and unsupervised video object segmentation.

AK

305,683 Aufrufe • vor 3 Jahren

Tutorial: ComfyUI Mattes & Utility Nodes Tutorial for VFX In this video, VFX artist Doug Hogan walks through a ComfyUI workflow for generating production-ready utility passes directly from your footage. No manual rotoscoping required. What's covered: → Prompt-based background removal (RMBG) for fast, simple shots → SAM3 text-prompt segmentation for moving subjects and medium-complexity scenes → SAM3 point-based editor for precise matte control (think: crypto mattes for live action) → Face segmentation for beauty and skin work in comp → Stable depth and normals passes for relighting and CG integration → App Mode — a simplified UI for running all options without touching the node graph All outputs are tied to your source plate's native resolution and frame rate via get/set nodes, so the workflow adapts automatically.
3:01

Tutorial: ComfyUI Mattes & Utility Nodes Tutorial for VFX In this video, VFX artist Doug Hogan walks through a ComfyUI workflow for generating production-ready utility passes directly from your footage. No manual rotoscoping required. What's covered: → Prompt-based background removal (RMBG) for fast, simple shots → SAM3 text-prompt segmentation for moving subjects and medium-complexity scenes → SAM3 point-based editor for precise matte control (think: crypto mattes for live action) → Face segmentation for beauty and skin work in comp → Stable depth and normals passes for relighting and CG integration → App Mode — a simplified UI for running all options without touching the node graph All outputs are tied to your source plate's native resolution and frame rate via get/set nodes, so the workflow adapts automatically.

ComfyUI

95,385 Aufrufe • vor 1 Monat

Tripo Segmentation now breaks 3D models into meaningful parts based on their geometry. Get cleaner boundaries, more accurate part separation, and fewer manual fixes when you need production-ready results. Choose from Simple, Balanced, or Detailed to match your workflow. Perfect for 3D printing, game assets, and downstream editing.
0:24

Tripo Segmentation now breaks 3D models into meaningful parts based on their geometry. Get cleaner boundaries, more accurate part separation, and fewer manual fixes when you need production-ready results. Choose from Simple, Balanced, or Detailed to match your workflow. Perfect for 3D printing, game assets, and downstream editing.

Tripo 🔜 gamescom

25,897 Aufrufe • vor 2 Monaten

SAM 2 from Meta FAIR is the first unified model for real-time, promptable object segmentation in images & videos. Using the model in our web-based demo you can segment, track and apply effects to objects in video in just a few clicks. Try SAM 2 ➡️
0:40

SAM 2 from Meta FAIR is the first unified model for real-time, promptable object segmentation in images & videos. Using the model in our web-based demo you can segment, track and apply effects to objects in video in just a few clicks. Try SAM 2 ➡️

AI at Meta

88,918 Aufrufe • vor 2 Jahren

Mattes and Utilities Demo: VFX artist Doug Hogan breaks down a ComfyUI workflow that generates production-ready utility passes directly from your footage. No manual rotoscoping. No guesswork. What's inside: → Prompt-based background removal (RMBG) for fast, clean shots → SAM3 text-prompt segmentation for moving subjects → SAM3 point-based editor for precise matte control → Face segmentation for beauty and skin work in comp → Stable depth and normals passes for relighting and CG integration → App Mode - a simplified UI that hides the node graph entirely Everything ties back to your source plate's native resolution and frame rate automatically via get/set nodes. This is the kind of workflow that used to take a team. Now it runs in ComfyUI. Click the link below for the workflow 👇
3:01

Mattes and Utilities Demo: VFX artist Doug Hogan breaks down a ComfyUI workflow that generates production-ready utility passes directly from your footage. No manual rotoscoping. No guesswork. What's inside: → Prompt-based background removal (RMBG) for fast, clean shots → SAM3 text-prompt segmentation for moving subjects → SAM3 point-based editor for precise matte control → Face segmentation for beauty and skin work in comp → Stable depth and normals passes for relighting and CG integration → App Mode - a simplified UI that hides the node graph entirely Everything ties back to your source plate's native resolution and frame rate automatically via get/set nodes. This is the kind of workflow that used to take a team. Now it runs in ComfyUI. Click the link below for the workflow 👇

ComfyUI

26,631 Aufrufe • vor 2 Monaten

Wow, this is really cool. An app that will turn your Figma files into a browsable documentation site based on your library structure. The free version looks good enough to start with ✅ Stevie from FigMayo
0:41

Wow, this is really cool. An app that will turn your Figma files into a browsable documentation site based on your library structure. The free version looks good enough to start with ✅ Stevie from FigMayo

Ozan Öztaskiran

41,243 Aufrufe • vor 2 Jahren

A 1-billion-parameter motion model trained on NVIDIA CUDA, requiring 8GB of VRAM for real-time operation. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.
0:41

A 1-billion-parameter motion model trained on NVIDIA CUDA, requiring 8GB of VRAM for real-time operation. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

CYANPUPPETS

22,931 Aufrufe • vor 9 Monaten

AI FOOTBALL ANALYSIS. A FULL COMPUTER VISION SYSTEM. BUILT ON YOLO, OPENCV, AND PYTHON. You upload a regular match video. No sensors, no GPS trackers, just camera footage. The neural network finds every player, referee, and ball on its own. Every frame, in real time. KMeans clustering breaks down jersey colors pixel by pixel. The system splits players into teams automatically. Without a single manual hint. Optical Flow tracks camera movement. Separates it from player movement. Perspective Transformation converts pixels into real meters. Speed of every player. Distance covered. Ball possession percentage. All calculated automatically. Four hours of tutorial from zero to a working system. The model is trained on real Bundesliga matches. Runs on a regular GPU. Python code - take it and run. Sports analytics is no longer behind closed doors. AI leveled the playing field.
4:28:35

AI FOOTBALL ANALYSIS. A FULL COMPUTER VISION SYSTEM. BUILT ON YOLO, OPENCV, AND PYTHON. You upload a regular match video. No sensors, no GPS trackers, just camera footage. The neural network finds every player, referee, and ball on its own. Every frame, in real time. KMeans clustering breaks down jersey colors pixel by pixel. The system splits players into teams automatically. Without a single manual hint. Optical Flow tracks camera movement. Separates it from player movement. Perspective Transformation converts pixels into real meters. Speed of every player. Distance covered. Ball possession percentage. All calculated automatically. Four hours of tutorial from zero to a working system. The model is trained on real Bundesliga matches. Runs on a regular GPU. Python code - take it and run. Sports analytics is no longer behind closed doors. AI leveled the playing field.

zostaff

1,075,589 Aufrufe • vor 4 Monaten

asked Fable to improve my rendering of the hairy guy, this looks good enough! took just a few prompts in Omma AI to get it right, the hair material is just a cheap #tsl LambertNodeMaterial with some custom specular lighting model based on the Kajiya-Kay lobes
0:39

asked Fable to improve my rendering of the hairy guy, this looks good enough! took just a few prompts in Omma AI to get it right, the hair material is just a cheap #tsl LambertNodeMaterial with some custom specular lighting model based on the Kajiya-Kay lobes

Marco Fugaro

39,932 Aufrufe • vor 1 Monat

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.
0:30

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

CYANPUPPETS

18,317 Aufrufe • vor 8 Monaten

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.
0:32

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

CYANPUPPETS

32,379 Aufrufe • vor 9 Monaten

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.
0:39

Blender AI Motion Plugin — a 1-billion-parameter motion model trained on NVIDIA CUDA, running locally in real time with 8GB of VRAM. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

CYANPUPPETS

125,040 Aufrufe • vor 8 Monaten

This video was based on a real conversation I had with a car hacker who told me that sometimes all you need in order to hack into a car is its license plate. And now the government has mandated kill switches in cars. Good luck securing those from hackers.
0:55

This video was based on a real conversation I had with a car hacker who told me that sometimes all you need in order to hack into a car is its license plate. And now the government has mandated kill switches in cars. Good luck securing those from hackers.

Naomi Brockwell priv/acc

15,189 Aufrufe • vor 2 Monaten

Arena AutoEval can quickly evaluate new models, allowing model labs to leverage results faster. It compresses the evaluation feedback loop from weeks to hours, enabling rapid, human-preference-based insight into a model’s capabilities. For more details, see the full video in the post below.
1:35

Arena AutoEval can quickly evaluate new models, allowing model labs to leverage results faster. It compresses the evaluation feedback loop from weeks to hours, enabling rapid, human-preference-based insight into a model’s capabilities. For more details, see the full video in the post below.

Arena.ai

21,381 Aufrufe • vor 17 Tagen

STEPS: – Go to ChatGPT and write a situation based on your idea. – Ask it to refine the situation, make any necessary edits, and turn it into a complete story. – Ask ChatGPT to create a 30-second scenario from your story. – Then ask ChatGPT to turn that scenario into a 12–15 frame storyboard. – Upload the storyboard to ChatGPT and ask it to turn the storyboard into a detailed video prompt. – Finally, upload the storyboard to Seedance, copy the video prompt, and paste it into Seedance. – Within 5–6 minutes, you can have a complete video based on your original idea. P.S. Like, repost, and bookmark! . . .
0:52

STEPS: – Go to ChatGPT and write a situation based on your idea. – Ask it to refine the situation, make any necessary edits, and turn it into a complete story. – Ask ChatGPT to create a 30-second scenario from your story. – Then ask ChatGPT to turn that scenario into a 12–15 frame storyboard. – Upload the storyboard to ChatGPT and ask it to turn the storyboard into a detailed video prompt. – Finally, upload the storyboard to Seedance, copy the video prompt, and paste it into Seedance. – Within 5–6 minutes, you can have a complete video based on your original idea. P.S. Like, repost, and bookmark! . . .

Beginnersblog

15,623 Aufrufe • vor 10 Tagen

Is Apple's creativity officially tapped out? We asked 𝐀𝐠𝐞𝐧𝐭 𝐎𝐩𝐮𝐬 to "Create a clear, professional news-style explainer video about the latest #AppleEvent." In less than an hour, it produced a video that looks like something you'd expect from a newsroom. With 𝐀𝐠𝐞𝐧𝐭 𝐎𝐩𝐮𝐬, anyone can turn headline news into a professional and accurate video in real time with AI. 👉 Replay link:
0:55

Is Apple's creativity officially tapped out? We asked 𝐀𝐠𝐞𝐧𝐭 𝐎𝐩𝐮𝐬 to "Create a clear, professional news-style explainer video about the latest #AppleEvent." In less than an hour, it produced a video that looks like something you'd expect from a newsroom. With 𝐀𝐠𝐞𝐧𝐭 𝐎𝐩𝐮𝐬, anyone can turn headline news into a professional and accurate video in real time with AI. 👉 Replay link:

Agent Opus

1,062,463 Aufrufe • vor 11 Monaten

Been using Claude 3.7 for a couple days and it is very, very good. Its "vibe coding" from language is impressive. I wrote a whole Substack post about the model (& Grok 3), link in reply. Here is a one-shot prompted video game based on the Melville story “Bartleby the Scrivner”
0:47

Been using Claude 3.7 for a couple days and it is very, very good. Its "vibe coding" from language is impressive. I wrote a whole Substack post about the model (& Grok 3), link in reply. Here is a one-shot prompted video game based on the Melville story “Bartleby the Scrivner”

Ethan Mollick

155,684 Aufrufe • vor 1 Jahr