https://github.com/mike10004/requery-face

Java port of Liu's CCV face detection
https://github.com/mike10004/requery-face

face-detection java

Last synced: 15 days ago
JSON representation

Java port of Liu's CCV face detection

• Host: GitHub
• URL: https://github.com/mike10004/requery-face
• Owner: mike10004
• License: bsd-3-clause
• Created: 2017-06-22T21:30:41.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2017-06-26T19:45:09.000Z (over 7 years ago)
• Last Synced: 2024-10-07T09:21:24.889Z (3 months ago)
• Topics: face-detection, java
• Language: JavaScript
• Size: 1.37 MB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# requery-face

This is a pure Java port of [Liu Liu's JavaScript face detection code][liuliu]. 

Currently, it's a pretty quick and dirty job. It is *not* any of these things:

* fast

* robust

* memory-conserving

But if what you're looking for is Java face detection, and you're not all that

concerned with speed or detecting all faces, then this may be for you. Liu Liu

did fine work in writing C code and JavaScript code for face detection, and 

you should check those out if you're really interested in high performance.

## Usage

For an example of how to use the code, check out the `Annotate` class, found

among the unit tests. It is a program that runs over a directory of JPEG files

and re-writes them with face annotations.

The gist of it is something like this:

    BufferedImage image = ImageIO.read(imageFile);

    RequeryFaceDetector detector = new RequeryFaceDetector();

    List detections = detector.detect(BufferedCanvas.from(image));

It has never been tested with anything but the default set of classifiers and 

the default detection options. Use at your own risk. 

## To do

Some suggestions on how to improve the code:

* use single-channel representations of images (instead of 4-channel RGBA)

* parallelize the image data analysis part 

[liuliu]: http://libccv.org/