Overview
What this challenge is about.
Use a small provided dataset of around 200 part images under 3 lighting conditions. Build a classical pipeline using OpenCV: grayscale + adaptive thresholding + Canny edge detection + morphological cleanup + burr counting via connected components. Evaluate burr-count agreement with the human-annotated ground truth (mean absolute error). Compare in writing (not code) against a hypothetical deep-learning approach on cost, data needs, and explainability. Deliver the pipeline, an evaluation notebook, and a 3-page memo for the quality-engineering team.
The Brief
What you'll do, and what you'll demonstrate.
Build a lighting-robust classical edge-detection pipeline that counts burrs within ±1 of human-annotated ground truth and document when a deep-learning approach would be worth the upgrade.
Earning criteria — what you'll demonstrate
- Apply classical image-processing operations to a real QA problem
- Reason about when classical methods beat deep learning
- Evaluate a vision pipeline against human-annotated ground truth
- Document a vision pipeline so it survives lighting changes
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
Classical CV pipelines on real factory data are a daily portion of CV-engineer work in manufacturing-AI roles.
This challenge sharpens
- image-processing
- edge-detection
- opencv
AI Engineer
Reasoning about when classical methods are sufficient (instead of jumping to deep learning) is the AI-engineer judgment that saves teams months of work.
This challenge sharpens
- image-processing
- documentation
- opencv
Machine Learning Engineer
Robust evaluation against ground truth across operating conditions is the MLE habit production teams rely on.
This challenge sharpens
- evaluation
- opencv
- edge-detection