A Girls Guide To Growing A Moustache with Python
summary
A pipeline that finds a donor moustache and transposes it onto another face using OpenCV and dlib runs on a Raspberry Pi 3; it detects faces with a landmark model trained on ~60,000 images, trains a custom moustache classifier with ~500 positive and thousands of negative samples, and composes results using image pyramids, histogram of oriented gradients and Delaunay triangles plus Procrustes alignment and blending. Results are demonstrable on live images but require more positive training for cleaner seams and lighting.
AI-written summary · about
chapters (13)
- 0:00Intro
- 0:41Pythoness
- 1:50The Moustache
- 8:36Edges
- 10:00Sober Edge Detector
- 10:49Facial Features
- 11:31Aligning
- 11:56Triangles
- 12:32Blending
- 14:20Pro Crusties
- 14:53Training
- 15:28Samples
- 15:45Variations
related talks
29mThe Mustache Movement
21mBuilding your own dystopic surveillance state with Python
37mUnstructured Data Processing with a Raspberry Pi AI Kit and Python
39mBuilding Highly Scalable Facial Recognition Pipelines
01h 12mBuilding a live face recognition system
30mThe Art and Science of Deep Face Recognition in Python
40mScience Art in the Cloud: How I built a Museum Exhibit
21mLeveraging Python for Real-Time Image processing