What is this?
An entry-level Python project for face, video and text detection and recognition. It bundles 11 demo features, from face detection in photos and video to contour marking, hat overlays, digital makeup, gender and emotion recognition, image inpainting and auto colorization.
An entry-level Python project for face, video and text detection and recognition. It bundles 11 demo features, from face detection in photos and video to contour marking, hat overlays, digital makeup, gender and emotion recognition, image inpainting and auto colorization.Is this an app I install, or code I run myself?It is a collection of Python demo scripts plus step-by-step tutorials, not an installable product. You clone the repo and run the scripts for each feature in your own Python environment.
It is a collection of Python demo scripts plus step-by-step tutorials, not an installable product. You clone the repo and run the scripts for each feature in your own Python environment.What do I need to install to run the demos?A fairly old pinned stack: Windows 10 x64, Python 3.6.4, OpenCV 3.4.1, Dlib 19.8.1, facerecognition 1.2.2, keras 2.1.6, tensorflow 1.8.0 and Tesseract OCR 4.0.0-beta.1. The README links setup tutorials, starting with the OpenCV environment guide.
A fairly old pinned stack: Windows 10 x64, Python 3.6.4, OpenCV 3.4.1, Dlib 19.8.1, facerecognition 1.2.2, keras 2.1.6, tensorflow 1.8.0 and Tesseract OCR 4.0.0-beta.1. The README links setup tutorials, starting with the OpenCV environment guide.Which libraries power each feature?Face recognition uses OpenCV or Dlib, face detection uses facerecognition, gender recognition runs on keras plus tensorflow, and text recognition uses Tesseract OCR.
Face recognition uses OpenCV or Dlib, face detection uses facerecognition, gender recognition runs on keras plus tensorflow, and text recognition uses Tesseract OCR.Are all eleven features actually finished?No. Eye tracking and face swap are explicitly marked as work in progress, and the TODO list also mentions Dlib performance optimization and model-training guides as open items.
No. Eye tracking and face swap are explicitly marked as work in progress, and the TODO list also mentions Dlib performance optimization and model-training guides as open items.I'm new to computer vision — is this a good place to learn?Yes, that is the intent: the author positions it as entry-level, and every feature comes with a linked tutorial and preview images, so you can study each capability in isolation rather than fighting a big framework.
Yes, that is the intent: the author positions it as entry-level, and every feature comes with a linked tutorial and preview images, so you can study each capability in isolation rather than fighting a big framework.What can it actually do with a face?It detects and recognizes faces in photos and video, draws the 68 facial contour points, overlays effects like hats, applies digital makeup (lipstick, eyebrows, eyes), and recognizes gender plus seven emotions: angry, disgust, fear, happy, sad, surprise and neutral.
It detects and recognizes faces in photos and video, draws the 68 facial contour points, overlays effects like hats, applies digital makeup (lipstick, eyebrows, eyes), and recognizes gender plus seven emotions: angry, disgust, fear, happy, sad, surprise and neutral.Does it work with video, not just photos?Yes — face detection and recognition both run on video, with separate OpenCV and Dlib versions, and there is also video object extraction for pulling a subject out of footage.
Yes — face detection and recognition both run on video, with separate OpenCV and Dlib versions, and there is also video object extraction for pulling a subject out of footage.What else can it do besides faces?Text recognition via Tesseract OCR, image repair that can be used for watermark removal, and automatic colorization of photos. Eye tracking and face swap are still incomplete.
Text recognition via Tesseract OCR, image repair that can be used for watermark removal, and automatic colorization of photos. Eye tracking and face swap are still incomplete.Is the documentation available in English?The README is written in Chinese, but it links an 'English Doc' at the top, and the feature previews are images, so most demos are understandable either way.
The README is written in Chinese, but it links an 'English Doc' at the top, and the feature previews are images, so most demos are understandable either way.What are similar projects?X-AnyLabeling, Ai-Learn, deepface, face-api.js, facenet, face_recognition, learnopencv, video-subtitle-remover
X-AnyLabeling, Ai-Learn, deepface, face-api.js, facenet, face_recognition, learnopencv, video-subtitle-removerWhich topics is it in?emotion-recognition, watermark-removal, opencv, face-recognition, face-detection, keras, tensorflow, video-processing, ocr, image-processing, computer-vision, tutorial
emotion-recognition, watermark-removal, opencv, face-recognition, face-detection, keras, tensorflow, video-processing, ocr, image-processing, computer-vision, tutorial