libfacedetection

所属分类:图形图象
开发工具:C/C++
文件大小:8021KB
下载次数:17
上传日期:2017-12-27 18:46:55
上 传 者sin1731
说明:  多角度人脸识别算法,采用自定义库,识别速度大于opencv。
(Multi angle face recognition algorithm.)

文件列表:
bin (0, 2017-12-21)
bin\libfacedetect-x64.dll (9960448, 2017-12-04)
bin\libfacedetect.dll (9787904, 2017-12-04)
ChangeLog (1414, 2017-12-04)
example (0, 2017-12-27)
example\benchmark.cpp (4581, 2017-12-04)
example\libfacedetect-example.cpp (7051, 2017-12-04)
FDDB-results-of-4functions.png (53862, 2017-12-04)
images (0, 2017-12-21)
images\chloecalmon.png (790085, 2017-12-04)
images\keliamoniz1.jpg (64424, 2017-12-04)
images\keliamoniz2.jpg (88213, 2017-12-04)
include (0, 2017-12-21)
include\facedetect-dll.h (4905, 2017-12-04)
lib (0, 2017-12-21)
lib\libfacedetect-x64.lib (3246, 2017-12-04)
lib\libfacedetect.lib (3134, 2017-12-04)
LICENSE (1100, 2017-12-04)

# libfacedetection This is a binary library for face detection and face landmark detection in images. The 32-bit and ***-bit dll files are provided. To achieve better performance, the ***-bit dll is recommended. examples/libfacedetect-example.cpp shows how to use the library. ![Examples](/images/chloecalmon.png "Detection example") Comparison on Windows ------------- | Method | Time | FPS |Time | FPS |Time | FPS | Misc | |--------------------|-------------|-------------|-------------|-------------|-------------|------------|--------| | | Win32 | Win32 | X*** | X*** | X*** |X*** | | | |Single-thread|Single-thread|Single-thread|Single-thread|Multi-thread |Multi-thread| | |OpenCV | -- | -- | -- | -- | 12.33ms | 81.1 | Yaw angle: -60 to 60 degrees| |frontal | 2.92ms | 342.5 | 2.41ms | 414.9 | 0.652ms | 1533.1 | Yaw angle: -60 to 60 degrees| |frontal-surveillance| 3.83ms | 261.1 | 3.37ms | 269.7 | 0.944ms | 1059.8 | Yaw angle: -70 to 70 degrees | |multiview | 7.12ms | 140.4 | 5.81ms | 172.1 | 1.597ms | 62*** | Yaw angle: -90 to 90 degrees | |multiview_reinforce | 10.95ms | 91.3 | 9.15ms | 109.3 | 2.725ms | 367.0 | Yaw angle: -90 to 90 degrees | * Face detection only, and no landmark detection included. * ***0x480 image size (VGA), scale=1.2, minimal window size = 48. * Intel(R) Core(TM) i7-4770 CPU @ 3.4GHz. * OpenCV classifier file: haarcascade_frontalface_alt.xml Comparison on ARM ------------- | Method | Time | FPS | Misc | |--------------------|--------|------|--------| |frontal | 12.5ms| 80.0 | Yaw angle: -60 to 60 degrees| |frontal-surveillance| 15.7ms| 63.7 | Yaw angle: -70 to 70 degrees | |multiview | 27.8ms| 36.0 | Yaw angle: -90 to 90 degrees | |multiview_reinforce | 38.4ms| 26.0 | Yaw angle: -90 to 90 degrees | * Face detection only, and no landmark detection included. * ***0x480 image size (VGA), scale=1.2, minimal window size = 48 * NVIDIA TK1 "4-Plus-1" 2.32GHz ARM quad-core Cortex-A15 CPU * Multi-core parallelization is disabled. * C programming language, and no SIMD instruction is used. 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. Binary library for ARM ------------- The binary evaluation library for ARM can be downloaded at https://github.com/OAID/YSQfastfd . The detection functions can only be called about 2000 times for evaluation. Accuracy evaluation ------------- FDDB: http://vis-www.cs.umass.edu/fddb/index.html ![Evaluation on FDDB](https://github.com/ShiqiYu/libfacedetection/blob/master/FDDB-results-of-4functions.png "Evaluation on FDDB") * scale=1.08 * minimal window size = 16 * the heights of the face rectangles are scaled to 1.2 to fit the ground truth data in FDDB. Author ------------- * Shiqi Yu, Contributors ------------- * Jia Wu * Shengyin Wu * Dong Xu ------------- * The result image was taken by [Chlo Calmon](https://www.instagram.com/chloecalmon/).

近期下载者

相关文件


收藏者