Chiebot-Mirror
/
opencv_contrib
8
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Make detector interface simpler

Browse Source
pull/21/head
Vlad Shakhuro 11 years ago
parent cf08b72646
commit 0d97598754
4 changed files with 39 additions and 354 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 42
      modules/adas/include/opencv2/adas.hpp
  2. 111
      modules/adas/include/opencv2/adas/acffeature.hpp
  3. 105
      modules/adas/include/opencv2/adas/icfdetector.hpp
  4. 135
      modules/adas/include/opencv2/adas/waldboost.hpp

42
modules/adas/include/opencv2/adas.hpp
Unescape View File

@ -42,8 +42,44 @@ the use of this software, even if advised of the possibility of such damage.
#ifndef __OPENCV_ADAS_ADAS_HPP__
#define __OPENCV_ADAS_ADAS_HPP__
#include "adas/acffeature.hpp"
#include "adas/waldboost.hpp"
#include "adas/icfdetector.hpp"
namespace cv
{
namespace adas
{
class CV_EXPORTS ICFDetector
{
public:
/* Initialize detector */
ICFDetector();
/* Load detector from file, return true on success, false otherwise */
bool load(const std::string& filename);
/* Run detector on single image
image — image for detection
bboxes — output array of object bounding boxes in format
Rect(row_from, col_from, n_rows, n_cols)
confidence_values — output array of confidence values from 0 to 1.
One value per bbox — confidence of detector that corresponding
bbox contatins object
minObjSize — min possible object size on image in pixels (rows x cols)
maxObjSize — max possible object size on image in pixels (rows x cols)
*/
void detect(InputArray image,
std::vector<Rect>& bboxes,
std::vector<float>& confidence_values,
Size minObjSize,
Size maxObjSize) const;
};
} /* namespace adas */
} /* namespace cv */
#endif /* __OPENCV_ADAS_ADAS_HPP__ */

111
modules/adas/include/opencv2/adas/acffeature.hpp
Unescape View File

@ -1,111 +0,0 @@
/*
By downloading, copying, installing or using the software you agree to this
license. If you do not agree to this license, do not download, install,
copy or use the software.
License Agreement
For Open Source Computer Vision Library
(3-clause BSD License)
Copyright (C) 2013, OpenCV Foundation, all rights reserved.
Third party copyrights are property of their respective owners.
Redistribution and use in source and binary forms, with or without modification,
are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the documentation
and/or other materials provided with the distribution.
* Neither the names of the copyright holders nor the names of the contributors
may be used to endorse or promote products derived from this software
without specific prior written permission.
This software is provided by the copyright holders and contributors "as is" and
any express or implied warranties, including, but not limited to, the implied
warranties of merchantability and fitness for a particular purpose are
disclaimed. In no event shall copyright holders or contributors be liable for
any direct, indirect, incidental, special, exemplary, or consequential damages
(including, but not limited to, procurement of substitute goods or services;
loss of use, data, or profits; or business interruption) however caused
and on any theory of liability, whether in contract, strict liability,
or tort (including negligence or otherwise) arising in any way out of
the use of this software, even if advised of the possibility of such damage.
*/
#ifndef __OPENCV_ADAS_ACFFEATURE_HPP__
#define __OPENCV_ADAS_ACFFEATURE_HPP__
namespace cv
{
namespace adas
{
/* Compute channel pyramid for acf features
image — image, for which pyramid should be computed
params — pyramid computing parameters
Returns computed channels in vectors N x CH,
N — number of scales (outer vector),
CH — number of channels (inner vectors)
*/
std::vector<std::vector<Mat_<int>>>
computeChannels(const Mat& image, const ScaleParams& params);
class ACFFeatureEvaluator
{
public:
/* Construct evaluator, set features to evaluate */
ACFFeatureEvaluator(const std::vector<Point>& features);
/* Set channels for feature evaluation */
void setChannels(const std::vector<Mat_<int>>& channels);
/* Set window position */
void setPosition(Size position);
/* Evaluate feature with given index for current channels
and window position */
int evaluate(size_t feature_ind) const;
/* Evaluate all features for current channels and window position
Returns matrix-column of features
*/
Mat_<int> evaluateAll() const;
private:
/* Features to evaluate */
std::vector<Point> features_;
/* Channels for feature evaluation */
std::vector<Mat_<int>> channels
/* Channels window position */
Size position_;
};
/* Generate acf features
window_size — size of window in which features should be evaluated
count — number of features to generate.
Max number of features is min(count, # possible distinct features)
seed — random number generator initializer
Returns vector of distinct acf features
*/
std::vector<Point>
generateFeatures(Size window_size, size_t count = UINT_MAX, int seed = 0);
} /* namespace adas */
} /* namespace cv */
#endif /* __OPENCV_ADAS_ACFFEATURE_HPP__ */

105
modules/adas/include/opencv2/adas/icfdetector.hpp
Unescape View File

@ -1,105 +0,0 @@
/*
By downloading, copying, installing or using the software you agree to this
license. If you do not agree to this license, do not download, install,
copy or use the software.
License Agreement
For Open Source Computer Vision Library
(3-clause BSD License)
Copyright (C) 2013, OpenCV Foundation, all rights reserved.
Third party copyrights are property of their respective owners.
Redistribution and use in source and binary forms, with or without modification,
are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the documentation
and/or other materials provided with the distribution.
* Neither the names of the copyright holders nor the names of the contributors
may be used to endorse or promote products derived from this software
without specific prior written permission.
This software is provided by the copyright holders and contributors "as is" and
any express or implied warranties, including, but not limited to, the implied
warranties of merchantability and fitness for a particular purpose are
disclaimed. In no event shall copyright holders or contributors be liable for
any direct, indirect, incidental, special, exemplary, or consequential damages
(including, but not limited to, procurement of substitute goods or services;
loss of use, data, or profits; or business interruption) however caused
and on any theory of liability, whether in contract, strict liability,
or tort (including negligence or otherwise) arising in any way out of
the use of this software, even if advised of the possibility of such damage.
*/
#ifndef __OPENCV_ADAS_ICFDETECTOR_HPP__
#define __OPENCV_ADAS_ICFDETECTOR_HPP__
namespace cv
{
namespace adas
{
class ICFDetector
{
public:
/* Initialize detector
min_obj_size — min possible object size on image in pixels (rows x cols)
max_obj_size — max possible object size on image in pixels (rows x cols)
scales_per_octave — number of images in pyramid while going
from scale x to scale 2x. Affects on speed
and quality of the detector
*/
ICFDetector(Size min_obj_size,
Size max_obj_size,
int scales_per_octave = 8);
/* Load detector from file, return true on success, false otherwise */
bool load(const std::string& filename);
/* Run detector on single image
image — image for detection
bboxes — output array of bounding boxes in format
Rect(row_from, col_from, n_rows, n_cols)
confidence_values — output array of confidence values from 0 to 1.
One value per bbox — confidence of detector that corresponding
bbox contatins object
*/
void detect(InputArray image,
std::vector<Rect>& bboxes,
std::vector<float>& confidence_values) const;
/* Train detector
image_filenames — filenames of images for training
labelling — vector of object bounding boxes per every image
params — parameters for detector training
*/
void train(const std::vector<std::string>& image_filenames,
const std::vector<std::vector<Rect>>& labelling,
const ICFDetectorParams& params = ICFDetectorParams());
/* Save detector in file, return true on success, false otherwise */
bool save(const std::string& filename);
};
} /* namespace adas */
} /* namespace cv */
#endif /* __OPENCV_ADAS_ICFDETECTOR_HPP__ */

135
modules/adas/include/opencv2/adas/waldboost.hpp
Unescape View File

@ -1,135 +0,0 @@
/*
By downloading, copying, installing or using the software you agree to this
license. If you do not agree to this license, do not download, install,
copy or use the software.
License Agreement
For Open Source Computer Vision Library
(3-clause BSD License)
Copyright (C) 2013, OpenCV Foundation, all rights reserved.
Third party copyrights are property of their respective owners.
Redistribution and use in source and binary forms, with or without modification,
are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the documentation
and/or other materials provided with the distribution.
* Neither the names of the copyright holders nor the names of the contributors
may be used to endorse or promote products derived from this software
without specific prior written permission.
This software is provided by the copyright holders and contributors "as is" and
any express or implied warranties, including, but not limited to, the implied
warranties of merchantability and fitness for a particular purpose are
disclaimed. In no event shall copyright holders or contributors be liable for
any direct, indirect, incidental, special, exemplary, or consequential damages
(including, but not limited to, procurement of substitute goods or services;
loss of use, data, or profits; or business interruption) however caused
and on any theory of liability, whether in contract, strict liability,
or tort (including negligence or otherwise) arising in any way out of
the use of this software, even if advised of the possibility of such damage.
*/
#ifndef __OPENCV_ADAS_WALDBOOST_HPP__
#define __OPENCV_ADAS_WALDBOOST_HPP__
namespace cv
{
namespace adas
{
class Stump
{
public:
/* Train stump for given data
data — matrix of feature values, size M x N, one feature per row
labels — matrix of sample class labels, size 1 x N. Labels can be from
{-1, +1}
Returns chosen feature index. Feature enumeration starts from 0
*/
int train(const Mat& data, const Mat& labels);
/* Predict object class given
value — feature value. Feature must be the same as chose during training
stump
Returns object class from {-1, +1}
*/
int predict(int value);
private:
/* Stump decision threshold */
int threshold_;
/* Stump polarity, can be from {-1, +1} */
int polarity_;
/* Stump decision rule:
h(value) = polarity * sign(value - threshold)
*/
};
/* Save Stump to FileStorage */
FileStorage& operator<< (FileStorage& out, const Stump& classifier);
/* Load Stump from FileStorage */
FileStorage& operator>> (FileStorage& in, Stump& classifier);
class WaldBoost
{
public:
/* Initialize WaldBoost cascade with default of specified parameters */
WaldBoost(const WaldBoostParams& = WaldBoostParams());
/* Train WaldBoost cascade for given data
data — matrix of feature values, size M x N, one feature per row
labels — matrix of sample class labels, size 1 x N. Labels can be from
{-1, +1}
Returns feature indices chosen for cascade.
Feature enumeration starts from 0
*/
std::vector<int> train(const Mat& data,
const Mat& labels);
/* Predict object class given object that can compute object features
feature_evaluator — object that can compute features by demand
Returns confidence_value — measure of confidense that object
is from class +1
*/
float predict(const Ptr<ACFFeatureEvaluator>& feature_evaluator);
private:
/* Parameters for cascade training */
WaldBoostParams params_;
/* Stumps in cascade */
std::vector<Stump> stumps_;
/* Weight for stumps in cascade linear combination */
std::vector<float> stump_weights_;
/* Rejection thresholds for linear combination at every stump evaluation */
std::vector<float> thresholds_;
};
/* Save WaldBoost to FileStorage */
FileStorage& operator<< (FileStorage& out, const WaldBoost& classifier);
/* Load WaldBoost from FileStorage */
FileStorage& operator>> (FileStorage& in, WaldBoost& classifier);
} /* namespace adas */
} /* namespace cv */
#endif /* __OPENCV_ADAS_WALDBOOST_HPP__ */
Write Preview
Loading…
Cancel
Save