Overview
What this challenge is about.
Use COLMAP (open-source SfM) + OpenMVS (open-source MVS) on a curated dataset of 5 small artifacts plus a calibration cube. Build a single Python CLI that ingests a folder of images and outputs a textured mesh + a metric-accuracy report against the cube's known 10 cm edge. Target sub-1mm mean error on the cube. Deliver the pipeline, evaluation, and a 4-page productization memo (per-artifact cost, time-to-deliver, quality tier).
The Brief
What you'll do, and what you'll demonstrate.
Build a reproducible photo-to-textured-mesh pipeline with sub-1mm metric error on a calibration cube and a per-artifact cost story.
Earning criteria — what you'll demonstrate
- Run a full SfM + MVS pipeline on a real small-object dataset
- Apply a calibration target for metric validation
- Wrap research-grade tools in a productizable CLI
- Communicate productization trade-offs to a non-engineering founder
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
Computer Vision Engineer
Full SfM + MVS pipelines on small objects are CV-engineer portfolio work at any AR, mapping, or heritage-tech company.
This challenge sharpens
- structure-from-motion
- multi-view-stereo
- 3d-reconstruction
Applied AI Scientist
Metric validation against a known calibration target is the applied-AI rigor that distinguishes shippable from demo.
This challenge sharpens
- geometric-validation
- 3d-reconstruction
- mesh-generation
AI Engineer
Productizing a research pipeline as a single-command CLI is the AI-engineer-as-toolsmith role that startups depend on.
This challenge sharpens
- python
- 3d-reconstruction
- mesh-generation