Online 2D Photo → 3D Point Cloud Converter (AI Depth + OBJ Export) Budget: $150–$200 I’m looking for someone who can build a web-based tool that converts standard 2D photos into 3D point cloud / micro-triangle style models using AI depth estimation — similar to effects used for 3D engraved crystal keepsakes. This is the foundation of a future product line — so clean, reliable code is important. Main Features Upload 2D photo (PNG/JPG no background) AI depth map generation Convert to 3D point cloud (bubblegram / micro triangles) Adjustable density (approx 0.03–0.07) Export .OBJ In-browser 3D preview if possible Runs on my Ubuntu server: • Mac Pro 2013 with AMD GPU Technical Requirements Python (NumPy, OpenCV, PCL, PyTorch/ONNX) Depth model experience: Depth Anything V2, ZoeDepth, MiDaS, etc. OBJ export + geometry generation Web-based UI: FastAPI or Flask backend (PHP-friendly helpful) Clean HTML/CSS/JS frontend Micro-triangle mesh generation GPU optimization Basic noise cleanup / background toggle Deliverables Fully working online app hosted on my server Upload → Convert → Preview → Download Adjustable UI controls: Point density Point size Depth scale Optional background removal toggle Complete source code + setup documentation Why This Project Is Fun & Meaningful You’ll be helping create a system that turns flat photos into volumetric art — floating points inside crystal cubes! Your work will be used in: Personalized 3D portrait laser engraving Unique 3D memory preservation This blends: AI 3D graphics Real-world manufacturing …and there’s a big roadmap ahead with more work for the right person: Improved face enhancement Advanced noise cleanup Faster GPU inference pipeline Direct engraving path output If you like seeing code create something magical — this is for you Tags / Skills AI • 3D Point Cloud • OBJ Export • Python • OpenCV • NumPy • PyTorch • Depth Anything V2 • ZoeDepth • MiDaS • ONNX • PCL • 3D Geometry • Web App • FastAPI • Flask • Ubuntu Server • GPU Optimization • 3D Visualization • Three.js • Computer Vision I Will Provide Example input images Reference OBJ files Server credentials after NDA Visual target example (video) If you're excited by innovative 3D + AI experiences, let's build something awesome together! Please include relevant samples of similar AI/3D work when applying.