A fast binary library for face detection in images. The face detection speed can reach 1400FPS. You can use it free of charge with any purpose.
Switch branches/tags
Nothing to show
Clone or download
Latest commit b84b69f Nov 17, 2018
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
bin Removed old AdaBoost methods Nov 17, 2018
example Removed old AdaBoost methods Nov 17, 2018
images Update the example image. Oct 23, 2018
include Removed old AdaBoost methods Nov 17, 2018
lib Removed old AdaBoost methods Nov 17, 2018
ChangeLog Removed old AdaBoost methods Nov 17, 2018
FDDB-results-of-4functions.png speed up 2x to 3x & parallelization is disabled Oct 6, 2016
LICENSE Speedup 1.2x Jun 8, 2016
README.md minor rev. Nov 17, 2018

README.md

libfacedetection

This is a binary library for CNN-based face detection in images.

examples/libfacedetectcnn-example.cpp shows how to use the library.

Examples

CNN-based Face Detection on Windows

Method Time FPS Time FPS
X64 X64 X64 X64
Single-thread Single-thread Multi-thread Multi-thread
OpenCV Haar+AdaBoost (640x480) -- -- 12.33ms 81.1
cnn (CPU, 640x480) 69.03ms 14.49 16.47ms 60.72
cnn (CPU, 320x240) 16.54ms 60.46 4.15ms 241.00
cnn (CPU, 160x120) 3.79ms 263.65 1.01ms 989.98
cnn (CPU, 128x96) 2.53ms 395.29 0.71ms 1399.28
  • OpenCV Haar+AdaBoost runs with minimal face size 48x48
  • Face detection only, and no landmark detection included.
  • Minimal face size ~10x10
  • Intel(R) Core(TM) i7-7700 CPU @ 3.6GHz.

CNN-based Face Detection on ARM Linux (Raspberry Pi 3 B+)

Method Time FPS Time FPS
Single-thread Single-thread Multi-thread Multi-thread
cnn (CPU, 640x480) 593.86ms 1.68 183.85ms 5.44
cnn (CPU, 320x240) 140.50ms 7.12 45.48ms 21.99
cnn (CPU, 160x120) 30.15ms 33.17 10.75ms 92.99
cnn (CPU, 128x96) 20.20ms 49.49 6.73ms 148.53
  • Face detection only, and no landmark detection included.
  • Minimal face size ~10x10
  • Raspberry Pi 3 B+, Broadcom BCM2837B0, Cortex-A53 (ARMv8) 64-bit SoC @ 1.4GHz

The dll cannot run on ARM. The library should be recompiled from source code for ARM compatibility. If you need the source code, a commercial license is needed.

Author

Contributors

  • Jia Wu
  • Shengyin Wu
  • Dong Xu