一张照片就能生成3D模型，MIT开源随便用，导出的文件可以直接导入Blender。 TripoSplat是TripoAI开源的一个工具，输入任意一张2D图片，它就能输出高质量的3D高斯泼溅模型。不止静态预览，导出的.ply和.splat文件可以在SparkJS、SuperSplat里直接打开编辑，也能导入ComfyUI工作流。 不需要账号，不需要付费，部署在自己的机器上运行也行。 使用教程： 方式一，在线Demo，零安装 打开HuggingFace上的TripoSplat Space，上传图片就能生成3D模型，浏览器直接预览。 方式二，本地安装 1. 下载模型权重：pip install huggingface_hub，然后用python脚本下载VAST-AI/TripoSplat到ckpts目录 2. 安装依赖：pip install numpy safetensors pillow tqdm torch torchvision 3. 运行：python run_example.py 如果有GPU推理速度会快很多，CPU也能运行就是慢一点。 如果报CUDA错误的话，先确认torch版本和CUDA版本匹配，pip list | grep torch查一下。 github.com/VAST-AI-Research/…

The scenes coming out of The Latent Space (a collaboration with Alejandro Cartagena / @halecar2 ) are now becoming explorable world fragments. Using the LoRAs trained on Alejandro's datasets, paper compositions on my desk are passed through the realtime Diffusion pipeline, then turned into 3D Gaussian Splats using ML Sharp.⁣ ⁣ #diffsuion #GaussianSplatting #MLSharp #SparkJS #AI

Yes, this version is 27mb .spz file type. Then using the level of detail (LoD) in sparkjs can get this to (sorta) run on quest hah

【完全解説】 AIで3D空間に"意味ラベル"を自動付与するデモがすごい x.com/bilawalsidhu/status/20… これ、Gemini 3.1 SparkJSで 3D Gaussian Splatの中をスキャンすると、 「これは椅子」 「これは植物」 「これは壁」 みたいな意味情報を、3D空間上にそのまま固定できるやつ👀 流れはこう👇 1. 任意の3Dシーンを読み込んでスキャン 2. Geminiに画面内の物体を2D検出させる 3. カメラを少し動かしながら複数視点で認識 4. 検出結果をクラスタリング 5. 3D空間上の正しい位置に投影 6. 永続的な「3Dセマンティックレイヤー」として保存 つまり、 ただの3Dスキャンデータが "AIが理解できる空間データ"に変わる。 これが進むとかなりデカい。 ・ARグラスが現実世界の物体を理解する ・ロボットが部屋の中の意味を把握する ・3Dマップ内で「机の上の本」みたいに検索できる ・ゲーム/VR空間に後付けで意味情報を付与できる 今まで3Dデータは「見た目の再現」が中心だったけど、 ここからは「AIが読める3D空間」になっていく。 地味に見えて、AR・ロボティクス・空間AIの土台になりそう。 他の海外テックの話題はこちら👇 @trendtech33566

Gemini 3.1 SparkJS を使用して、3D ガウススプラットをリアルタイムで意味論的に注釈付け 1. 任意の 3D シーンを読み込み、スキャンボタンを押す 2. VLM から 2D 検出を取得 3. 出力をクラスタリングし、3D ワールド空間に投影 4. 永続的な 3D 意味論レイヤーとして保存

Gemini 3.1とSparkJSを使い、3DGSの街や建物のデータに自動でタグ付けする試み。 AIの2D認識結果を3D空間へ計算して落とし込み、タグ付けした位置情報と名前がそのまま残すことができる模様。 2D向けの視覚AIを3Dに応用した、興味深い実験です。

Semantically annotating 3D gaussian splats on the fly using gemini 3.1 sparkjs 1. Load any 3D scene and hit scan 2. Get 2D detections from VLM 3. Cluster outputs & project into 3D world space 4. Save as a persistent 3D semantic layer Inspired by @alexanderchen's experiments with gemini visual intelligence. Just had to try to lift it from 2D to 3D!

Gaussian splats are good at places. Games need levels. Tested @theworldlabs Marble inside Bevy (ported sparkjs → github.com/htdt/bevy_spark). I’m not putting it into Godogen yet. It makes nice walkable spaces, but I need controllable game levels.

A sample of the world fragments from yesterday's Hands-on Hallucination session. Physical interfaces → real-time img2img diffusion → 3D GS → real-time rendering in a browser on desktop, mobile and AVP.⁣ #diffsuion #GaussianSplatting #MLSharp #SparkJS #AI

Accelerated capture from yesterday's Hands-on Hallucination session. I started out testing the system alone, then my son joined in and quickly took control; deciding which generated frames to freeze and transform into 3D Gaussian Splats. We ended up with a series of strange little world fragments.⁣⁣ ⁣ Physical interfaces → real-time img2img diffusion → 3D GS → real-time rendering in a browser on desktop, mobile and AVP.⁣ ⁣⁣ #GaussianSplatting #AppleML #MLSharp #SparkJS

I resumed work on my Hands-on Hallucination ML Sharp experiments from ~5 months ago. I manipulate paper, clay, etc. in front of a camera, watch a real-time diffusion pipeline hallucinate a world from them, press a knob, and the frozen frame becomes a 3D Gaussian Splat, instantly explorable in the browser or in VR. #GaussianSplatting #AppleML #MLSharp #SparkJS

Dormant Memories — a gallery regrouping my recent experiments built from real-world captures, reconstructed as 3D Gaussian Splats, and in many cases manipulated using @theworldlabs' Marble. smallfly.com/dormant_memorie… Everything is built with Spark, the 3D Gaussian splat renderer released by World Labs. It runs in the browser and can be experienced across devices: desktop, mobile, and headsets like Apple Vision Pro and Quest 3. This work comes out of something I've been exploring for a few years now: capturing fragments of the world (places, objects, moments...) and turning them into spaces that can then be revisited, manipulated, and explored in real time. What particularly interests me here is the process of creating alternate versions of the same place; starting from something that physically exists and using it as a base to explore variations, instead of generating from scratch. These explorations are part of what we do at @bonjourdpt... same tools, same curiosity, same collaborative energy that shape the studio's work on interactive and immersive experiences. Some of these experiments were featured as part of the Spark 2.0 release by World Labs. Their technical blog post is a must-read; link in the thread ↓ These are experiments. They aren't fully optimized, they aren't bug-free, and the UX has rough edges. The goal is to explore, test, and learn. #GaussianSplatting #3DGS #pointcloud #WorldModels #sparkjs #threejs

Thoughts on playcanvas vs sparkjs?

Release: LichtFeld Studio v0.5.2 is out! With 186 commits to master, this release is mainly focused on bug fixes and usability improvements. What’s new in v0.5.2: - Histogram cleanup improvements - HDR image maps for rendering - Background color picker - Visualization of loss per camera - Training evaluation now also works in the UI - Free orbit and FPV camera modes - Dockable sequencer and histogram - Continued migration from ImGui to RmlUi for HTML- and CSS-like formatting and styling This release is available to all supporters as a Windows binary via portal.lichtfeld.io. At the same time, LichtFeld Studio remains committed to being free and open source under GPLv3, and it can also be built from source. Please consider supporting the ongoing development of LichtFeld Studio through a donation via the portal. That support has already enabled the first grant to Shady, who will contribute significantly to the upcoming v0.5.3 release. Thank you to everyone who supports this project financially, with code, bug reports, data, website work, and other contributions. With continued support, we hope to expand these grants to support more active contributors to the codebase. A special thank you to our foundational sponsor Core11 and our gold sponsor Volinga, who have helped make the current state of the software possible. Thank you as well to every donor and to all the new bronze sponsors. For v0.5.3, LFS will move the frontend to Vulkan, which is the first step toward vendor-independent GPU support. It will also enable initial path tracing and custom shaders. Planned features also include direct support for the sparkjs rad LoD format and asset management for training data and 3DGS files. You can find links to our main sponsors in the comments. Please visit also the website to explore our new bronze sponsors! Hint: We do not yet have a silver or platinum sponsor. 😉

Spark 2.0 で追加されたパラメータを検証用アプリを作って動作確認中...🧪 LoD 設定がとてもいい感じに機能している様子なので、まもなく Threedive AI にも反映予定！ #SparkJS #3DGS #ThreediveAI

Starting from the real world, capture it, and use it as a base to explore variations, instead of generating from scratch. Here capturing this alley with the PortalCam, then point cloud to 3D Gaussian splats and finally manipulated with Marble/@theworldlabs. The result is the same space across multiple alternate realities — abandoned, decayed, day, night — connected through portals. Everything runs real time in the browser using Spark 2.0/@sparkjsdev. #GaussianSplatting #3DGS #pointcloud #WorldModels #sparkjs #threejs

Oh I don't use an engine. I use sparkjs / threejs for 3D, and pixijs directly for 2D.

Third-person rail shooter. 3DGS, SparkJS

Web 3D 的新紀元！Spark 2.0 正式把百萬級場景搬進瀏覽器 🚀 一直以來，在網頁端流暢渲染大規模 3D 真實場景都是個挑戰。現在，Spark 2.0 帶著開源力量來解決這個問題了！ 基於 Three.js WebGL2 原生開發，這項技術達成了幾個驚人的指標： • 100M Splat 巨型世界： 支援流式加載（Streamable LoD），不用等下載完就能開始看。 • 跨設備流暢度： 從手機、電腦到 VR 設備都能絲滑運行，不再是高配電腦的專利。 • 真實世界還原： 透過 Gaussian Splatting，在瀏覽器就能感受百萬級像素的真實感。 這絕對是 Web 3D 開發者的必看項目。 #SparkJS #Threejs #Web3D #GaussianSplatting #開源 #WebGL2