Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

This open-source tool generates normal maps for pixel art sprites It works with both single sprites and animated sprite sheets, turning flat pixel art into assets that can react to dynamic lighting in-game. You can run it: > locally from the CLI > directly in the browser with WASM... show more

Hakimi Eiqbal

2,379 subscribers

300,719 Aufrufe • vor 2 Tagen •via X (Twitter)

Kunst Wissenschaft & Technologie

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

The Fortnite Sprite Season is almost ending. How we feel on them? 🟥 How many sprites did you manage to collect - master? Do you like this system? MY TIPS: 🔰PERK COMBO - only 1 perk per sprite makes them extremely uninteresting gameplay wise. And turned sprites into nothing but a collection Don't make it so i must have the sprite on my back for its perk. Allow me to combo 3 perks ch4 augments style without them being on my back. As the sprites you extract unlock you their perks to combo! 🔰BONUS PERK - If you combo 3 sprites of the same version, ex: cube. You get the special bonus perk of that sprite version on too! So 4 perks total. Like combo 3 gummy's for extra sprite dust. 🔰Make it so having sprites on your back is exclusive for leveling them up and extracting. Or showing off even if someone wants to. How limited sprites are has made them so useless gameplay wise I think If Fortnite makes these changes, Sprites go from just collect them onto generational gameplay potential! ❤️ Let me know what you think of this

The Fortnite Sprite Season is almost ending. How we feel on them? 🟥 How many sprites did you manage to collect - master? Do you like this system? MY TIPS: 🔰PERK COMBO - only 1 perk per sprite makes them extremely uninteresting gameplay wise. And turned sprites into nothing but a collection Don't make it so i must have the sprite on my back for its perk. Allow me to combo 3 perks ch4 augments style without them being on my back. As the sprites you extract unlock you their perks to combo! 🔰BONUS PERK - If you combo 3 sprites of the same version, ex: cube. You get the special bonus perk of that sprite version on too! So 4 perks total. Like combo 3 gummy's for extra sprite dust. 🔰Make it so having sprites on your back is exclusive for leveling them up and extracting. Or showing off even if someone wants to. How limited sprites are has made them so useless gameplay wise I think If Fortnite makes these changes, Sprites go from just collect them onto generational gameplay potential! ❤️ Let me know what you think of this

Scyan

26,049 Aufrufe • vor 2 Monaten

Can't stop thinking about the art style Desimulate created so I tried to create it in Three.js to see if I could achieve a similar look and feel. I think I got pretty close with some careful pixel art billboarding and lighting. But there's some limitations to getting this particular perspective right so it would only work in environments where you limit where the player can walk to. Vibe coded on Grok 4, Claude Sonnet using Cursor.

Can't stop thinking about the art style Desimulate created so I tried to create it in Three.js to see if I could achieve a similar look and feel. I think I got pretty close with some careful pixel art billboarding and lighting. But there's some limitations to getting this particular perspective right so it would only work in environments where you limit where the player can walk to. Vibe coded on Grok 4, Claude Sonnet using Cursor.

Danny Limanseta

232,509 Aufrufe • vor 1 Jahr

you are making video games all wrong for decades almost all game art time went into reverse-engineering the end result you start with concept art of what you WANT it to look like. then for YEARS you walk the computer backwards into something it can process: a rendering engine, 3d assets that fit that engine, and a gamer’s local hardware that has to turn all of that back into a high fidelity image incredibly slow. incredibly expensive. but it was the best we had ai makes that pipeline faster as everyone is figuring out today. but they’re missing the real unlock: YOU DON’T NEED THAT PIPELINE AT ALL if you have a reference for the end look, you don’t model it in 3d. you don’t reverse-engineer into an engine you keep a very basic greybox model of the world as the source of truth, and track every entity in real time. it doesn’t need to look good at all… the player never sees it! a fast video model turns that into the “game” in real time. it doesn’t have to look like a game. it can look like a 2d cartoon or cinematic live action fal’s h3 max is very fast. astra 6 low-effort too. neither is fully realtime yet but together they can generate continuous, coherent, player-controlled scenes, faster than playback early demo:

you are making video games all wrong for decades almost all game art time went into reverse-engineering the end result you start with concept art of what you WANT it to look like. then for YEARS you walk the computer backwards into something it can process: a rendering engine, 3d assets that fit that engine, and a gamer’s local hardware that has to turn all of that back into a high fidelity image incredibly slow. incredibly expensive. but it was the best we had ai makes that pipeline faster as everyone is figuring out today. but they’re missing the real unlock: YOU DON’T NEED THAT PIPELINE AT ALL if you have a reference for the end look, you don’t model it in 3d. you don’t reverse-engineer into an engine you keep a very basic greybox model of the world as the source of truth, and track every entity in real time. it doesn’t need to look good at all… the player never sees it! a fast video model turns that into the “game” in real time. it doesn’t have to look like a game. it can look like a 2d cartoon or cinematic live action fal’s h3 max is very fast. astra 6 low-effort too. neither is fully realtime yet but together they can generate continuous, coherent, player-controlled scenes, faster than playback early demo:

Udi Wertheimer

125,114 Aufrufe • vor 1 Monat

IF YOU SEE THIS POST, IT MEANS I'M TOO BROKE TO SAY NO TO NEW PROJECTS 🥲 Hi, I'm Nera, an art director leading a small pixel art team. We've just wrapped up our latest project, and now my wallet has started asking uncomfortable questions. So... we're looking for new commissions or a game project to join. We can help with pretty much everything on the art side: • Characters • Environments • Props • Animations • UI • Art direction Whether you're an experienced developer or someone building your very first game, we'd love to hear about it. Help us make cool games before my bank account starts rendering in low resolution.🥲🥲 DMs are open. ❤️

IF YOU SEE THIS POST, IT MEANS I'M TOO BROKE TO SAY NO TO NEW PROJECTS 🥲 Hi, I'm Nera, an art director leading a small pixel art team. We've just wrapped up our latest project, and now my wallet has started asking uncomfortable questions. So... we're looking for new commissions or a game project to join. We can help with pretty much everything on the art side: • Characters • Environments • Props • Animations • UI • Art direction Whether you're an experienced developer or someone building your very first game, we'd love to hear about it. Help us make cool games before my bank account starts rendering in low resolution.🥲🥲 DMs are open. ❤️

Nera Nera - Pixel art game

14,254 Aufrufe • vor 2 Monaten

Pixel art was MADE to be viewed on a CRT TV. The blooming, scanlines, etc are part of the image that smooths it out, blends colors together, and gets rid of the harsh edges. The original game artists created everything with that analog distortion in mind, so the raw pixels aren't necessarily what they intended you to see. Don't get me wrong, raw pixels can look cool too! And if you like that look, then by all means go for it. But it's worth trying a nice CRT shader on them to see if you might like it even better!

Pixel art was MADE to be viewed on a CRT TV. The blooming, scanlines, etc are part of the image that smooths it out, blends colors together, and gets rid of the harsh edges. The original game artists created everything with that analog distortion in mind, so the raw pixels aren't necessarily what they intended you to see. Don't get me wrong, raw pixels can look cool too! And if you like that look, then by all means go for it. But it's worth trying a nice CRT shader on them to see if you might like it even better!

Finn McKenty

555,986 Aufrufe • vor 2 Monaten

Day 23 of how I vibe code my game engine in Cursor: Unit Sprites Like I mentioned in my buildings post: units hit the same wall once the animation steps outside the sprite’s inner silhouette. That’s where the pain starts: pixel bleeding, halos, and PBR maps becoming impossible to mask properly because they must follow the whole sprite. I almost gave in and embraced the demon lords: 3D → 2.5D. But then I had an idea: What if every PBR map (normal/emission/etc) could inherit the exact same animation as the diffuse sprite? If the motion matches 1:1, we get the full 2.5D look without breaking the animation, and we can run it at whatever framerate we want. So I built a prototype and… it actually works. (check the vid) Pretty neat, right? Next step: streamline it in-engine and scale it to all 6 hex movement directions without turning it into a manual nightmare. Also investigating the remaining white aura issue — I might regenerate diffuse + normal + emission in a 2×3 grid on both white + black backgrounds and extract cleaner alpha. After that: compositing layered units (ex: tank + turret). More pain ahead. But it’s the fun kind!

Day 23 of how I vibe code my game engine in Cursor: Unit Sprites Like I mentioned in my buildings post: units hit the same wall once the animation steps outside the sprite’s inner silhouette. That’s where the pain starts: pixel bleeding, halos, and PBR maps becoming impossible to mask properly because they must follow the whole sprite. I almost gave in and embraced the demon lords: 3D → 2.5D. But then I had an idea: What if every PBR map (normal/emission/etc) could inherit the exact same animation as the diffuse sprite? If the motion matches 1:1, we get the full 2.5D look without breaking the animation, and we can run it at whatever framerate we want. So I built a prototype and… it actually works. (check the vid) Pretty neat, right? Next step: streamline it in-engine and scale it to all 6 hex movement directions without turning it into a manual nightmare. Also investigating the remaining white aura issue — I might regenerate diffuse + normal + emission in a 2×3 grid on both white + black backgrounds and extract cleaner alpha. After that: compositing layered units (ex: tank + turret). More pain ahead. But it’s the fun kind!

Startracker 🔺

121,234 Aufrufe • vor 9 Monaten

IF YOU SEE THIS POST, IT MEANS I'M LOOKING FOR A GAME PROJECT TO JOIN Hi, I'm Nera, an art director leading a small game art team. We've recently finished our latest game project, and we're now looking for a new game to work on. Our team can help with the full art production of your game from characters, environments, objects, and animations to UI. We can also help shape the visual direction of your game from the ground up. You don't need to be an experienced developer or have a large team to make a game. If you have an idea you're passionate about, we'd love to hear about it. If you're working on a game and looking for an art team, feel free to DM me. 🥰🩷

IF YOU SEE THIS POST, IT MEANS I'M LOOKING FOR A GAME PROJECT TO JOIN Hi, I'm Nera, an art director leading a small game art team. We've recently finished our latest game project, and we're now looking for a new game to work on. Our team can help with the full art production of your game from characters, environments, objects, and animations to UI. We can also help shape the visual direction of your game from the ground up. You don't need to be an experienced developer or have a large team to make a game. If you have an idea you're passionate about, we'd love to hear about it. If you're working on a game and looking for an art team, feel free to DM me. 🥰🩷

Nera Nera - Pixel art game

19,804 Aufrufe • vor 3 Monaten

Businesses, you can now* accept payments directly from agents—without, or with, a human in the loop—using Stripe machine payments. It works for both cards and stablecoins via Machine Payments Protocol or x402. Add to your existing integration in a single prompt. *No more waitlist, get started =>

Businesses, you can now* accept payments directly from agents—without, or with, a human in the loop—using Stripe machine payments. It works for both cards and stablecoins via Machine Payments Protocol or x402. Add to your existing integration in a single prompt. *No more waitlist, get started =>

Jeff Weinstein

12,529 Aufrufe • vor 5 Monaten

Stop manually making flowcharts. Try this GitHub skill with Claude Code. I found an open-source tool that turns a plain description of a software system into a clean, interactive diagram. It's called Archify, and it has more than 66,000 stars on GitHub. You add it as a skill to an AI coding agent like Claude Code, Codex, Cursor, or OpenCode. Then you describe the system in the chat in one line. Archify turns that line into a single HTML file you can open, click through, and send to anyone. If you point your agent at an existing codebase, it reads the code and Archify maps how the pieces connect. I liked five things about it. 1. It makes five kinds of diagrams, including system maps and step-by-step workflows. 2. You can click any piece and trace what connects to it. 3. It checks the layout before it hands the diagram over, so arrows don't pile up and labels don't overlap. 4. It has dark and light themes and exports to PNG, SVG, or a short video. 5. You keep editing in plain chat, with requests like "add Redis" or "highlight the rollback path." You don't need a codebase to use it. A description in the chat is enough.

Stop manually making flowcharts. Try this GitHub skill with Claude Code. I found an open-source tool that turns a plain description of a software system into a clean, interactive diagram. It's called Archify, and it has more than 66,000 stars on GitHub. You add it as a skill to an AI coding agent like Claude Code, Codex, Cursor, or OpenCode. Then you describe the system in the chat in one line. Archify turns that line into a single HTML file you can open, click through, and send to anyone. If you point your agent at an existing codebase, it reads the code and Archify maps how the pieces connect. I liked five things about it. 1. It makes five kinds of diagrams, including system maps and step-by-step workflows. 2. You can click any piece and trace what connects to it. 3. It checks the layout before it hands the diagram over, so arrows don't pile up and labels don't overlap. 4. It has dark and light themes and exports to PNG, SVG, or a short video. 5. You keep editing in plain chat, with requests like "add Redis" or "highlight the rollback path." You don't need a codebase to use it. A description in the chat is enough.

Alex Veremeyenko

61,928 Aufrufe • vor 1 Monat

You can create split-stack videos like this in one click with the link below.👇 In the simple version, all you have to do is upload an image and it handles everything for you. But if you want more control, you can jump into graph mode and tweak each parameter to get more detailed, custom results.

You can create split-stack videos like this in one click with the link below.👇 In the simple version, all you have to do is upload an image and it handles everything for you. But if you want more control, you can jump into graph mode and tweak each parameter to get more detailed, custom results.

ComfyUI

28,215 Aufrufe • vor 6 Monaten

So, by now I'm sure you saw the AI fanart recreation of the Princess Mononoke trailer. It got a LOT of feedback, mostly from people telling the creator what a terrible thing he did and how AI assisted workflows can never be used for anything good or meaningful and most of all that it can never be used for art. That's such nonsense. Of course it can. If it's wielded by artists and great storytellers, great things will come of it. If you hand a violin to someone on the street they'll be able to make it make sounds.. if you hand it to a talented musician it becomes a conveyer of thoughts, emotion and art. Here's a small scene from Mononoke that I did an AI take on.. it's far from perfect and to animate this properly would require much more work. The AI workflows we have today are lacking so many control elements. But to think those things won't come.. come on. Of course they will. But most importantly, in a way where they don't overtake the art completely, but instead acts like an "amplifier" of the artists using the tools. This technology will be in most of our creative tools in the very near future with full controllability. Oh and I never liked the idea of turning old disney classics into live-action films, or doing the same with the beautiful work of Studio Ghibli. But for the purpose of this post I hope you'll forgive me ;-) #art #animation

So, by now I'm sure you saw the AI fanart recreation of the Princess Mononoke trailer. It got a LOT of feedback, mostly from people telling the creator what a terrible thing he did and how AI assisted workflows can never be used for anything good or meaningful and most of all that it can never be used for art. That's such nonsense. Of course it can. If it's wielded by artists and great storytellers, great things will come of it. If you hand a violin to someone on the street they'll be able to make it make sounds.. if you hand it to a talented musician it becomes a conveyer of thoughts, emotion and art. Here's a small scene from Mononoke that I did an AI take on.. it's far from perfect and to animate this properly would require much more work. The AI workflows we have today are lacking so many control elements. But to think those things won't come.. come on. Of course they will. But most importantly, in a way where they don't overtake the art completely, but instead acts like an "amplifier" of the artists using the tools. This technology will be in most of our creative tools in the very near future with full controllability. Oh and I never liked the idea of turning old disney classics into live-action films, or doing the same with the beautiful work of Studio Ghibli. But for the purpose of this post I hope you'll forgive me ;-) #art #animation

Martin Nebelong

18,377 Aufrufe • vor 2 Jahren

This week is already so hot. 🔥 Massive release from Decart : Lucy 2.0 a World Editing Model running at 1080p, 30FPS in realtime. This is truly exciting, the era of real-time generative reality is here. We are moving from watching AI video to living inside AI video. A breakthrough model capable of transforming the visual world in real-time. Moving beyond offline rendering, Lucy 2.0 delivers high-fidelity 1080p video generation with near-zero latency. Lucy 2.0 literally "redraws" the entire world pixel-by-pixel, while you are watching it. e.g. If you want to be an anime character, it doesn't just put a mask on you. It turns your skin into anime skin, your hair into anime hair, and the lighting in your room into anime lighting. Lucy 2.0 is also trained to stop the generated video from slowly falling apart over time, so the same stream can run much longer without faces and details drifting. So why is this a "Massive Deal"? Traditional AI video-generation model takes a prompt, you wait 10–20 minutes, and the computer "bakes" a video for you. You couldn't touch it or change it while it was happening. But Lucy 2.0 works like a mirror. It happens in real-time (30 frames per second). There is no waiting. You move your hand, the AI character moves its hand instantly. The craziest part isn't the visuals; it's the physics. Usually, AI hallucinations are glitchy—hands merge into faces, walls melt. Lucy 2.0 understands how the world works without being told. It knows that if you take off a helmet, there is hair underneath. It knows that if you splash water, droplets fly. It learned "physics" just by watching millions of videos. The physical behavior you see emerges from learned visual dynamics, not from engineered geometry or explicit physics engines. Their official technical report explicitly states that the model does not use traditional 3D engines, depth maps, or wireframes. It is a "pure diffusion model."

This week is already so hot. 🔥 Massive release from Decart : Lucy 2.0 a World Editing Model running at 1080p, 30FPS in realtime. This is truly exciting, the era of real-time generative reality is here. We are moving from watching AI video to living inside AI video. A breakthrough model capable of transforming the visual world in real-time. Moving beyond offline rendering, Lucy 2.0 delivers high-fidelity 1080p video generation with near-zero latency. Lucy 2.0 literally "redraws" the entire world pixel-by-pixel, while you are watching it. e.g. If you want to be an anime character, it doesn't just put a mask on you. It turns your skin into anime skin, your hair into anime hair, and the lighting in your room into anime lighting. Lucy 2.0 is also trained to stop the generated video from slowly falling apart over time, so the same stream can run much longer without faces and details drifting. So why is this a "Massive Deal"? Traditional AI video-generation model takes a prompt, you wait 10–20 minutes, and the computer "bakes" a video for you. You couldn't touch it or change it while it was happening. But Lucy 2.0 works like a mirror. It happens in real-time (30 frames per second). There is no waiting. You move your hand, the AI character moves its hand instantly. The craziest part isn't the visuals; it's the physics. Usually, AI hallucinations are glitchy—hands merge into faces, walls melt. Lucy 2.0 understands how the world works without being told. It knows that if you take off a helmet, there is hair underneath. It knows that if you splash water, droplets fly. It learned "physics" just by watching millions of videos. The physical behavior you see emerges from learned visual dynamics, not from engineered geometry or explicit physics engines. Their official technical report explicitly states that the model does not use traditional 3D engines, depth maps, or wireframes. It is a "pure diffusion model."

Rohan Paul

12,761 Aufrufe • vor 8 Monaten

Opus 5.5 found a pretty clever way to turn concept art into a playable 3D world The workflow starts with a cel-shaded village concept generated with GPT. Then: > Opus 5.5 recreates the layout with simple 3D blocks takes a screenshot from a fixed camera > GPT paints over those blocks using the concept art as reference > Opus builds a shader that projects the painting back onto the 3D geometry The result still looks like the original illustration, but it’s now a real 3D space you can walk around and enter. The same pipeline was repeated for interiors. Characters were generated through GPT Images and then animated with Minimax H3. This feels like a really interesting shortcut for getting a very specific 2D art direction into an actual playable game without manually modeling every detail.

Opus 5.5 found a pretty clever way to turn concept art into a playable 3D world The workflow starts with a cel-shaded village concept generated with GPT. Then: > Opus 5.5 recreates the layout with simple 3D blocks takes a screenshot from a fixed camera > GPT paints over those blocks using the concept art as reference > Opus builds a shader that projects the painting back onto the 3D geometry The result still looks like the original illustration, but it’s now a real 3D space you can walk around and enter. The same pipeline was repeated for interiors. Characters were generated through GPT Images and then animated with Minimax H3. This feels like a really interesting shortcut for getting a very specific 2D art direction into an actual playable game without manually modeling every detail.

Hakimi Eiqbal

58,362 Aufrufe • vor 5 Tagen

No way I'm sharing this for free I just found a tool that lets you drop a whole video template into Claude and it breaks it down for you clip by clip. then Claude can swap the character for a new one rewrite the entire script in your tone even make the video longer if you want. the crazy part is it does it all in one go from your conversation. RT + comment "tool" and I'll send you the link

No way I'm sharing this for free I just found a tool that lets you drop a whole video template into Claude and it breaks it down for you clip by clip. then Claude can swap the character for a new one rewrite the entire script in your tone even make the video longer if you want. the crazy part is it does it all in one go from your conversation. RT + comment "tool" and I'll send you the link

Ahmad

32,176 Aufrufe • vor 5 Monaten

What I love about Replit ⠕. Low floor to get started. Ultra high ceiling for what you can build. I told Agent to build a native C application that runs on my desktop without a web browser - AND compile it into JavaScript to run in the web browser. Here's a video showing both running

What I love about Replit ⠕. Low floor to get started. Ultra high ceiling for what you can build. I told Agent to build a native C application that runs on my desktop without a web browser - AND compile it into JavaScript to run in the web browser. Here's a video showing both running

jordwalke

43,773 Aufrufe • vor 8 Monaten

OpenClaw, but built for normal people. Sim is an open-source platform that lets you build AI agent workflows on a drag-and-drop canvas. Connect them to channels like Telegram and WhatsApp and deploy without writing a single line of code. They also have a built-in Copilot that generates entire workflows from plain English, which you can then tweak and customize in the UI. Key features: - Free and open-source (Apache 2.0) - Vector store integration for RAG-grounded agents - Self-host with one command (`npx simstudio`) - Run fully local with Ollama, no API keys needed - Supports vLLM for production-grade self-hosted inference The thing I really like about Sim is the level of control you get. You can add conditional branching, parallel execution, human-in-the-loop approval gates, and even nest workflows inside other workflows. Everything is visible on the canvas, so you know exactly what your agent is doing at every step. And you can build a workflow in Sim, deploy it as an MCP server, and plug it into any agent, including OpenClaw. I've shared the link to Sim's GitHub repo in the next tweet.

OpenClaw, but built for normal people. Sim is an open-source platform that lets you build AI agent workflows on a drag-and-drop canvas. Connect them to channels like Telegram and WhatsApp and deploy without writing a single line of code. They also have a built-in Copilot that generates entire workflows from plain English, which you can then tweak and customize in the UI. Key features: - Free and open-source (Apache 2.0) - Vector store integration for RAG-grounded agents - Self-host with one command (`npx simstudio`) - Run fully local with Ollama, no API keys needed - Supports vLLM for production-grade self-hosted inference The thing I really like about Sim is the level of control you get. You can add conditional branching, parallel execution, human-in-the-loop approval gates, and even nest workflows inside other workflows. Everything is visible on the canvas, so you know exactly what your agent is doing at every step. And you can build a workflow in Sim, deploy it as an MCP server, and plug it into any agent, including OpenClaw. I've shared the link to Sim's GitHub repo in the next tweet.

Akshay 🚀

52,426 Aufrufe • vor 7 Monaten

IntelliJ IDEA includes an open-source Java decompiler called Fernflower. But did you know that you can also use it as a standalone CLI program? And what's even better – it's now available on Homebrew! To install it, run the `brew install fernflower` command in the terminal. 🧵 Learn more about how to use Fernflower in the thread! 🧵

IntelliJ IDEA includes an open-source Java decompiler called Fernflower. But did you know that you can also use it as a standalone CLI program? And what's even better – it's now available on Homebrew! To install it, run the `brew install fernflower` command in the terminal. 🧵 Learn more about how to use Fernflower in the thread! 🧵

IntelliJ IDEA, a JetBrains IDE

13,184 Aufrufe • vor 1 Jahr

Some updates on the multiview vistadream pipeline with Rerun! Rerun came in extremely useful here, as being able to visualize depths at each stage of the pipeline allowed me to debug some nasty bugs. Since the last time, I was only working with a single image input. I've added in VGGT as my multiview pose + depth estimator. It works REALLY well for getting camera poses, but the depths are not that great. To try and fix that, I estimated depth maps from MoGeV2 for each of the views, and scale+shift aligned them so that they would match up to the confident sections of VGGT's depth predictions. You can see in the video just how much sharper the visualized 2d depth maps are! The biggest issue continues to be the multiview consistency 🫠 That's up next, along with actually training the Gaussian splat. Lots of work went into actually understanding inputs+outputs for VGGT. I had some funky bugs where the confidence values would all collapse to true I'm also really excited for this pipeline to use Difix3D+ Nvidia instead of Flux Inpainting, it seems like a better suited for a multiview pipeline.

Some updates on the multiview vistadream pipeline with Rerun! Rerun came in extremely useful here, as being able to visualize depths at each stage of the pipeline allowed me to debug some nasty bugs. Since the last time, I was only working with a single image input. I've added in VGGT as my multiview pose + depth estimator. It works REALLY well for getting camera poses, but the depths are not that great. To try and fix that, I estimated depth maps from MoGeV2 for each of the views, and scale+shift aligned them so that they would match up to the confident sections of VGGT's depth predictions. You can see in the video just how much sharper the visualized 2d depth maps are! The biggest issue continues to be the multiview consistency 🫠 That's up next, along with actually training the Gaussian splat. Lots of work went into actually understanding inputs+outputs for VGGT. I had some funky bugs where the confidence values would all collapse to true I'm also really excited for this pipeline to use Difix3D+ Nvidia instead of Flux Inpainting, it seems like a better suited for a multiview pipeline.

Pablo Vela

29,904 Aufrufe • vor 1 Jahr

THIS GUY VIBE CODED A BROWSER TOOL WITH CLAUDE CODE THAT DRAWS PROPERTY LINES ONTO DRONE FOOTAGE AND KEEPS THEM LOCKED TO THE GROUND AS THE DRONE MOVES real estate agents kept asking for those property boundary overlays on drone videos so buyers can see exactly where a lot ends. the only real option was after effects, which is a massive pro tool for one small repeatable job so he built a browser tool that does just that one thing > drop in a drone clip and draw the parcel outline once on the first frame > gpu motion tracking keeps that boundary stuck to the ground as the drone flies and the perspective shifts > if the line drifts anywhere, you drop a keyframe to correct it and it re-tracks the segments around it > style the line however you want, then export the finished video > free to use, you dont even need an account until you export he built one tool that solves one annoying job perfectly, for people who were never going to learn after effects anyway

THIS GUY VIBE CODED A BROWSER TOOL WITH CLAUDE CODE THAT DRAWS PROPERTY LINES ONTO DRONE FOOTAGE AND KEEPS THEM LOCKED TO THE GROUND AS THE DRONE MOVES real estate agents kept asking for those property boundary overlays on drone videos so buyers can see exactly where a lot ends. the only real option was after effects, which is a massive pro tool for one small repeatable job so he built a browser tool that does just that one thing > drop in a drone clip and draw the parcel outline once on the first frame > gpu motion tracking keeps that boundary stuck to the ground as the drone flies and the perspective shifts > if the line drifts anywhere, you drop a keyframe to correct it and it re-tracks the segments around it > style the line however you want, then export the finished video > free to use, you dont even need an account until you export he built one tool that solves one annoying job perfectly, for people who were never going to learn after effects anyway

Om Patel

404,227 Aufrufe • vor 3 Monaten