Learn how to detect and recognize faces in images or live video using OpenCV and Dlib.
Experiment with adding digital makeup effects like lip color or eyebrow tinting to portrait photos.
Classify a person's emotion from a photo using a Keras model trained on seven emotion categories.
Colorize a black-and-white image automatically using deep learning in Python.
Built with specific older versions of OpenCV, Dlib, Keras, and TensorFlow, running on current versions may require manual compatibility adjustments.
FaceAI is a beginner-oriented Python project that demonstrates a range of face and image processing techniques. The README is written in Chinese, and the description translates roughly to a starter-level collection covering face detection, recognition, and related visual effects. An English README is linked from the main file. The project covers about eleven capabilities. These include detecting faces in still images and video, drawing facial landmark outlines, adding hats or other overlays to a portrait, applying digital makeup such as lip color and eyebrow effects, recognizing a person's gender, and identifying one of seven emotions (anger, disgust, fear, happiness, sadness, surprise, and neutral). Additional features include extracting objects from video, restoring damaged areas of an image (which the README notes can also remove watermarks), automatically colorizing black-and-white images, and two unfinished features: eye tracking and face swapping. The technology behind each feature is spelled out: face recognition uses OpenCV and Dlib, face detection uses the face_recognition library, gender classification uses Keras and TensorFlow, and text recognition uses Tesseract OCR. The project was built and tested on Windows 10 with specific older versions of each library, so running it on a current setup may require adjustments. The README links to individual tutorial documents for each feature, covering topics like setting up the OpenCV environment, using Tesseract, detecting faces from images versus video using either OpenCV or Dlib, and applying various makeup or compositing effects. There is no formal license stated in the README. The project appears intended as a learning resource and personal project rather than production software, given the beginner framing and the two unfinished items still listed as TODO.
← vipstone on gitmyhub — every repo by this author, as a profile.
Verify against the repo before relying on details.