原文:https://blog.csdn.net/app_12062011/article/details/51422604
闲得没事,折腾下opencv 人脸识别,从样本制作到评估。
1.直接copy opencv里的源码,创建工程,添加opencv库,可以直接cmake源码,但我之所以自己创建工程,是想多学习,并且降低与源码框架的耦合度。
这里如果出现_imp__CreateToolbarEx 符号无法解析(error LNK2019: unresolved external symbol __imp__CreateToolbarEx@52 referenced in function "int __cdecl icvCreateTrackbar(char const *,char const *,int *,int,void (__cdecl*)(int),void (__cdecl*)(int,void *),void *)" (?icvCreateTrackbar@@YAHPBD0PAHHP6AXH@ZP6AXHPAX@Z3@Z))。则添加comctl32.lib 和vfw32.lib 即可。亲测。安装配置如下:
http://blog.sina.com.cn/s/blog_96ea53fb0101htfb.html
2.编译生成haartarining的lib和其他相关exe(如opencv_createsamples.exe)后,即可debug。
正样本制作:
对于正样本,通常的做法是先把所有正样本裁切好,并对尺寸做规整(即缩放至指定大小),由于HaarTraining训练时输入的正样本是vec文件,所以需要使用OpenCV自带的CreateSamples程序(在你所按照的opencv\bin下,如果没有需要编译opencv\apps\HaarTraining\make下的.dsw文件,注意要编译release版的)将准备好的正样本转换为vec文件。转换的步骤如下:
1) 制作一个正样本描述文件,用于描述正样本文件名(包括绝对路径或相对路径),正样本数目以及各正样本在图片中的位置和大小。典型的正样本描述文件如下:
posdata/1(10).bmp 1 1 1 23 23
posdata/1(11).bmp 1 1 1 23 23
posdata/1(12).bmp 1 1 1 23 23
不过你可以把描述文件放在你的posdata路径(即正样本路径)下,这样你就不需要加前面的相对路径了。同样它的生成方式可以用负样本描述文件的生成方法,最后用txt的替换工具将“bmp”全部替换成“bmp 1 1 1 23 23
”就可以了,如果你的样本图片多,用txt替换会导致程序未响应,你可以将内容拷到word下替换,然后再拷回来。bmp后面那五个数字分别表示图片个数,目标的起始位置及其宽高。这样就生成了正样本描述文件posdata.dat。
2) 运行CreateSamples程序。如果直接在VC环境下运行,可以在Project\Settings\Debug属性页的Program arguments栏设置运行参数。下面是一个运行参数示例:
-info D:\face\posdata\posdata.dat -vec D:\face\pos.vec -num 50 -w 20 -h 20
表示有50个样本,样本宽20,高20,正样本描述文件为posdata.dat,结果输出到pos.vec。
或者在dos下输入:
"D:\Program Files\OpenCV\bin\createsamples.exe" -info "posdata\posdata.dat" -vec data\pos.vec -num 50 -w 20 -h 20
运行完了会d:\face\data下生成一个*.vec的文件。该文件包含正样本数目,宽高以及所有样本图像数据。
Createsamples程序的命令行参数:
命令行参数:
-vec <vec_file_name>
训练好的正样本的输出文件名。
-img<image_file_name>
源目标图片(例如:一个公司图标)
-bg<background_file_name>
背景描述文件。
-num<number_of_samples>
要产生的正样本的数量,和正样本图片数目相同。
-bgcolor<background_color>
背景色(假定当前图片为灰度图)。背景色制定了透明色。对于压缩图片,颜色方差量由bgthresh参数来指定。则在bgcolor-bgthresh和bgcolor+bgthresh中间的像素被认为是透明的。
-bgthresh<background_color_threshold>
-inv
如果指定,颜色会反色
-randinv
如果指定,颜色会任意反色
-maxidev<max_intensity_deviation>
背景色最大的偏离度。
-maxangel<max_x_rotation_angle>
-maxangle<max_y_rotation_angle>,
-maxzangle<max_x_rotation_angle>
最大旋转角度,以弧度为单位。
-show
如果指定,每个样本会被显示出来,按下"esc"会关闭这一开关,即不显示样本图片,而创建过程继续。这是个有用的debug选项。
-w<sample_width>
输出样本的宽度(以像素为单位)
-h《sample_height》
输出样本的高度,以像素为单位。
opencv_createsamples.exe的参数
(createsamples.cpp)
[cpp] view plain copy print?
- " [-info <collection_file_name>]\n"
- " [-img <image_file_name>]\n"
- " [-vec <vec_file_name>]\n"
- " [-bg <background_file_name>]\n [-num <number_of_samples = %d>]\n"
- " [-bgcolor <background_color = %d>]\n"
- " [-inv] [-randinv] [-bgthresh <background_color_threshold = %d>]\n"
- " [-maxidev <max_intensity_deviation = %d>]\n"
- " [-maxxangle <max_x_rotation_angle = %f>]\n"
- " [-maxyangle <max_y_rotation_angle = %f>]\n"
- " [-maxzangle <max_z_rotation_angle = %f>]\n"
- " [-show [<scale = %f>]]\n"
- " [-w <sample_width = %d>]\n [-h <sample_height = %d>]\n"//默认24*24
-
" [-info <collection_file_name>]\n"
-
" [-img <image_file_name>]\n"
-
" [-vec <vec_file_name>]\n"
-
" [-bg <background_file_name>]\n [-num <number_of_samples = %d>]\n"
-
" [-bgcolor <background_color = %d>]\n"
-
" [-inv] [-randinv] [-bgthresh <background_color_threshold = %d>]\n"
-
" [-maxidev <max_intensity_deviation = %d>]\n"
-
" [-maxxangle <max_x_rotation_angle = %f>]\n"
-
" [-maxyangle <max_y_rotation_angle = %f>]\n"
-
" [-maxzangle <max_z_rotation_angle = %f>]\n"
-
" [-show [<scale = %f>]]\n"
-
" [-w <sample_width = %d>]\n [-h <sample_height = %d>]\n"//默认24*24
以下1)~4)是按顺序判断,且有且仅有一个
1)提供imagename 和vecname时,调用以下操作
[cpp] view plain copy print?
- /*
- * cvCreateTrainingSamples
- *
- * Create training samples applying random distortions to sample image and
- * store them in .vec file
- *
- * filename - .vec file name
- * imgfilename - sample image file name
- * bgcolor - background color for sample image
- * bgthreshold - background color threshold. Pixels those colors are in range
- * [bgcolor-bgthreshold, bgcolor+bgthreshold] are considered as transparent
- * bgfilename - background description file name. If not NULL samples
- * will be put on arbitrary background
- * count - desired number of samples
- * invert - if not 0 sample foreground pixels will be inverted
- * if invert == CV_RANDOM_INVERT then samples will be inverted randomly
- * maxintensitydev - desired max intensity deviation of foreground samples pixels
- * maxxangle - max rotation angles
- * maxyangle
- * maxzangle
- * showsamples - if not 0 samples will be shown
- * winwidth - desired samples width
- * winheight - desired samples height
- */
-
/*
-
* cvCreateTrainingSamples
-
*
-
* Create training samples applying random distortions to sample image and
-
* store them in .vec file
-
*
-
* filename - .vec file name
-
* imgfilename - sample image file name
-
* bgcolor - background color for sample image
-
* bgthreshold - background color threshold. Pixels those colors are in range
-
* [bgcolor-bgthreshold, bgcolor+bgthreshold] are considered as transparent
-
* bgfilename - background description file name. If not NULL samples
-
* will be put on arbitrary background
-
* count - desired number of samples
-
* invert - if not 0 sample foreground pixels will be inverted
-
* if invert == CV_RANDOM_INVERT then samples will be inverted randomly
-
* maxintensitydev - desired max intensity deviation of foreground samples pixels
-
* maxxangle - max rotation angles
-
* maxyangle
-
* maxzangle
-
* showsamples - if not 0 samples will be shown
-
* winwidth - desired samples width
-
* winheight - desired samples height
-
*/
2)提供imagename、bgfilename和infoname时
与1)类似
3)提供 infoname和 vecname时,调用以下操作(这里是我们训练需要的)
[cpp] view plain copy print?
- /*
- * cvCreateTrainingSamplesFromInfo
- *
- * Create training samples from a set of marked up images and store them into .vec file
- * infoname - file in which marked up image descriptions are stored
- * num - desired number of samples
- * showsamples - if not 0 samples will be shown
- * winwidth - sample width
- * winheight - sample height
- *
- * Return number of successfully created samples
- */
- int cvCreateTrainingSamplesFromInfo( const char* infoname, const char* vecfilename,
- int num,
- int showsamples,
- int winwidth, int winheight )
-
/*
-
* cvCreateTrainingSamplesFromInfo
-
*
-
* Create training samples from a set of marked up images and store them into .vec file
-
* infoname - file in which marked up image descriptions are stored
-
* num - desired number of samples
-
* showsamples - if not 0 samples will be shown
-
* winwidth - sample width
-
* winheight - sample height
-
*
-
* Return number of successfully created samples
-
*/
-
int cvCreateTrainingSamplesFromInfo( const char* infoname, const char* vecfilename,
-
int num,
-
int showsamples,
-
int winwidth, int winheight )
函数内容:读取当前图中所有标记的sample(x,y,w,h),并将其缩放到winwidth、winheight大小,故在这之前的人为缩放操作不需要
(可以看到,仅需要num、w、h参数)
4)仅vecname时,可以将vec里面的所有缩放后的samples都显示出来
[cpp] view plain copy print?
- /*
- * cvShowVecSamples
- *
- * Shows samples stored in .vec file
- *
- * filename
- * .vec file name
- * winwidth
- * sample width
- * winheight
- * sample height
- * scale
- * the scale each sample is adjusted to(这个scale与3中的缩放不是一回事,这里仅为了显示而再次缩放)
- */
- void cvShowVecSamples( const char* filename, int winwidth, int winheight, double scale );
-
/*
-
* cvShowVecSamples
-
*
-
* Shows samples stored in .vec file
-
*
-
* filename
-
* .vec file name
-
* winwidth
-
* sample width
-
* winheight
-
* sample height
-
* scale
-
* the scale each sample is adjusted to(这个scale与3中的缩放不是一回事,这里仅为了显示而再次缩放)
-
*/
-
void cvShowVecSamples( const char* filename, int winwidth, int winheight, double scale );
训练样本分为正例样本和反例样本,其中正例样本是指待检目标样本,反例样本指其它任意图片。
负样本
负样本可以来自于任意的图片,但这些图片不能包含目标特征。负样本由背景描述文件来描述。背景描述文件是一个文本文件,每一行包含了一个负样本图片的文件名(基于描述文件的相对路径)。该文件创建方法如下:
采用Dos命令生成样本描述文件。具体方法是在Dos下的进入你的图片目录,比如我的图片放在D:\face\posdata下,则:
按Ctrl+R打开Windows运行程序,输入cmd打开DOS命令窗口,输入d:回车,再输入cd D:\face\negdata进入图片路径,再次输入dir /b > negdata.dat,则会图片路径下生成一个negdata.dat文件,打开该文件将最后一行的negdata.dat删除,这样就生成了负样本描述文件。
样本制作完成,下面训练。
2、opencv_haartraining.exe的参数
(haartraining.cpp )
[cpp] view plain copy print?
- " -data <dir_name>\n"
- " -vec <vec_file_name>\n"
- " -bg <background_file_name>\n"
- " [-bg-vecfile]\n"
- " [-npos <number_of_positive_samples = %d>]\n"
- " [-nneg <number_of_negative_samples = %d>]\n"
- " [-nstages <number_of_stages = %d>]\n"
- " [-nsplits <number_of_splits = %d>]\n"
- " [-mem <memory_in_MB = %d>]\n"
- " [-sym (default)] [-nonsym]\n"
- " [-minhitrate <min_hit_rate = %f>]\n"
- " [-maxfalsealarm <max_false_alarm_rate = %f>]\n"
- " [-weighttrimming <weight_trimming = %f>]\n"
- " [-eqw]\n"
- " [-mode <BASIC (default) | CORE | ALL>]\n"
- " [-w <sample_width = %d>]\n"
- " [-h <sample_height = %d>]\n"
- " [-bt <DAB | RAB | LB | GAB (default)>]\n"
- " [-err <misclass (default) | gini | entropy>]\n"
- " [-maxtreesplits <max_number_of_splits_in_tree_cascade = %d>]\n"
- " [-minpos <min_number_of_positive_samples_per_cluster = %d>]\n"
-
" -data <dir_name>\n"
-
" -vec <vec_file_name>\n"
-
" -bg <background_file_name>\n"
-
" [-bg-vecfile]\n"
-
" [-npos <number_of_positive_samples = %d>]\n"
-
" [-nneg <number_of_negative_samples = %d>]\n"
-
" [-nstages <number_of_stages = %d>]\n"
-
" [-nsplits <number_of_splits = %d>]\n"
-
" [-mem <memory_in_MB = %d>]\n"
-
" [-sym (default)] [-nonsym]\n"
-
" [-minhitrate <min_hit_rate = %f>]\n"
-
" [-maxfalsealarm <max_false_alarm_rate = %f>]\n"
-
" [-weighttrimming <weight_trimming = %f>]\n"
-
" [-eqw]\n"
-
" [-mode <BASIC (default) | CORE | ALL>]\n"
-
" [-w <sample_width = %d>]\n"
-
" [-h <sample_height = %d>]\n"
-
" [-bt <DAB | RAB | LB | GAB (default)>]\n"
-
" [-err <misclass (default) | gini | entropy>]\n"
-
" [-maxtreesplits <max_number_of_splits_in_tree_cascade = %d>]\n"
-
" [-minpos <min_number_of_positive_samples_per_cluster = %d>]\n"
-
" -data <dir_name>\n"
-
" -vec <vec_file_name>\n"
-
" -bg <background_file_name>\n"
-
" [-bg-vecfile]\n"
-
" [-npos <number_of_positive_samples = %d>]\n"
-
" [-nneg <number_of_negative_samples = %d>]\n"
-
" [-nstages <number_of_stages = %d>]\n"
-
" [-nsplits <number_of_splits = %d>]\n"
-
" [-mem <memory_in_MB = %d>]\n"
-
" [-sym (default)] [-nonsym]\n"
-
" [-minhitrate <min_hit_rate = %f>]\n"
-
" [-maxfalsealarm <max_false_alarm_rate = %f>]\n"
-
" [-weighttrimming <weight_trimming = %f>]\n"
-
" [-eqw]\n"
-
" [-mode <BASIC (default) | CORE | ALL>]\n"
-
" [-w <sample_width = %d>]\n"
-
" [-h <sample_height = %d>]\n"
-
" [-bt <DAB | RAB | LB | GAB (default)>]\n"
-
" [-err <misclass (default) | gini | entropy>]\n"
-
" [-maxtreesplits <max_number_of_splits_in_tree_cascade = %d>]\n"
-
" [-minpos <min_number_of_positive_samples_per_cluster = %d>]\n"
样本创建之后,接下来要训练分类器,这个过程是由haartraining程序来实现的。该程序源码由OpenCV自带,且可执行程序在OpenCV安装目录的bin目录下。
Haartraining的命令行参数如下:
-data<dir_name>
存放训练好的分类器的路径名。
-vec<vec_file_name>
正样本文件名(由trainingssamples程序或者由其他的方法创建的)
-bg<background_file_name>
背景描述文件。
-npos<number_of_positive_samples>,
-nneg<number_of_negative_samples>
用来训练每一个分类器阶段的正/负样本。合理的值是:nPos = 7000;nNeg = 3000
-nstages<number_of_stages>
训练的阶段数。
-nsplits<number_of_splits>
决定用于阶段分类器的弱分类器。如果1,则一个简单的stump classifier被使用。如果是2或者更多,则带有number_of_splits个内部节点的CART分类器被使用。
-mem<memory_in_MB>
预先计算的以MB为单位的可用内存。内存越大则训练的速度越快。
-sym(default)
-nonsym
指定训练的目标对象是否垂直对称。垂直对称提高目标的训练速度。例如,正面部是垂直对称的。
-minhitrate《min_hit_rate》
每个阶段分类器需要的最小的命中率。总的命中率为min_hit_rate的number_of_stages次方。
-maxfalsealarm<max_false_alarm_rate>
没有阶段分类器的最大错误报警率。总的错误警告率为max_false_alarm_rate的number_of_stages次方。
-weighttrimming<weight_trimming>
指定是否使用权修正和使用多大的权修正。一个基本的选择是0.9
-eqw
-mode<basic(default)|core|all>
选择用来训练的haar特征集的种类。basic仅仅使用垂直特征。all使用垂直和45度角旋转特征。
-w《sample_width》
-h《sample_height》
训练样本的尺寸,(以像素为单位)。必须和训练样本创建的尺寸相同。
一个训练分类器的例子:
"D:\Program Files\OpenCV\bin\haartraining.exe" -data data\cascade -vec data\pos.vec -bg negdata\negdata.dat -npos 49 -nneg 49 -mem 200 -mode ALL -w 20 -h 20
训练结束后,会在目录data下生成一些子目录,即为训练好的分类器。
这一步需要用到performance.exe,该程序源码由OpenCV自带,且可执行程序在OpenCV安装目录的bin目录下。
performance.exe -data data/cascade -info posdata/test.dat -w 20 -h 20 -rs 30
performance的命令行参数如下:
Usage: ./performance
-data <classifier_directory_name>
-info <collection_file_name>
[-maxSizeDiff <max_size_difference = 1.500000>]
[-maxPosDiff <max_position_difference = 0.300000>]
[-sf <scale_factor = 1.200000>]
[-ni]
[-nos <number_of_stages = -1>]
[-rs <roc_size = 40>]
[-w <sample_width = 24>]
[-h <sample_height = 24>]
也可以用opencv的cvHaarDetectObjects函数进行检测:
CvSeq* faces = cvHaarDetectObjects( img, cascade, storage,1.1, 2, CV_HAAR_DO_CANNY_PRUNING,cvSize(40, 40) ); //3. 检测人脸
使用心得: