Contact person: Business Client

Phone:Show

Email:Show

Location: Fauske, Norway

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"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 • [login to view URL] • 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." (client-provided description)