Face detection in PHP
summary
Adds face detection to Rokka.io and hits reality: composer packages detect at most one face, PHP bindings for OpenCV are outdated, and cloud APIs miss small faces. Tests show Dilip gives the best local accuracy (23 faces detected out of 25 in one group photo). Builds a tiny C++ detector that prints coordinates, feeds those back into PHP to set focal crops, and reports that the solution is fast, repeatable, and deployable where C++ can be compiled.
AI-written summary · about
chapters (10)
- 0:00Intro
- 0:30What is face detection
- 1:30Introduction
- 2:42Opensource
- 3:12Why face detection
- 8:22OpenCV
- 12:25Dilib
- 16:04Deep learning
- 17:20Amazon Recognition
- 21:32Tealip vs OpenCV
related talks
57mFoundations of Facial Recognition
49mReal Time Face Detection through a Webcam Using AI
13mHow we used Face Detection to create a delightful User Experience
32mSelf-Portraits in JavaScript
22mHI and AI
41mOCR in the Wild World of Social Media
51mRealTime Exoskeleton-Mapping and Body Segmentation on the Browser with Tensorflow.js
36mClosing Keynote: Machine Learning APIs with PHP
