update documentation for softcascade module

apps/sft/fpool.cpp

@@ -240,14 +240,14 @@ void glob(const string& refRoot, const string& refExt, svector &refvecFiles)
 
 #endif
 
-// in the default case data folders should be alligned as following:
+// in the default case data folders should be aligned as following:
 // 1. positives: <train or test path>/octave_<octave number>/pos/*.png
 // 2. negatives: <train or test path>/octave_<octave number>/neg/*.png
 ScaledDataset::ScaledDataset(const string& path, const int oct)
 {
     dprintf("%s\n", "get dataset file names...");
 
-    dprintf("%s\n", "Positives globbing...");
+    dprintf("%s\n", "Positives globing...");
 
 #if !defined (_WIN32) && ! defined(__MINGW32__)
     glob(path + "/pos/octave_" + itoa(oct) + "/*.png", pos);
@@ -255,7 +255,7 @@ ScaledDataset::ScaledDataset(const string& path, const int oct)
     glob(path + "/pos/octave_" + itoa(oct),     "png", pos);
 #endif
 
-    dprintf("%s\n", "Negatives globbing...");
+    dprintf("%s\n", "Negatives globing...");
 #if !defined (_WIN32) && ! defined(__MINGW32__)
     glob(path + "/neg/octave_" + itoa(oct) + "/*.png", neg);
 #else

apps/sft/include/sft/config.hpp

@@ -93,7 +93,7 @@ struct Config
     // List of octaves for which have to be trained cascades (a list of powers of two)
     ivector octaves;
 
-    // Maximum number of positives that should be ised during training
+    // Maximum number of positives that should be used during training
     int positives;
 
     // Initial number of negatives used during training.
@@ -102,10 +102,10 @@ struct Config
     // Number of weak negatives to add each bootstrapping step.
     int btpNegatives;
 
-    // Inverse of scale for feature resazing
+    // Inverse of scale for feature resizing
     int shrinkage;
 
-    // Depth on weak classifier's desition tree
+    // Depth on weak classifier's decision tree
     int treeDepth;
 
     // Weak classifiers number in resulted cascade
@@ -120,10 +120,10 @@ struct Config
     // path to resulting cascade
     string outXmlPath;
 
-    // seed for fandom generation
+    // seed for random generation
     int seed;
 
-    // // bounding retangle for actual exemple into example window
+    // // bounding rectangle for actual example into example window
     // cv::Rect exampleWindow;
 };
 

apps/sft/sft.cpp

@@ -40,7 +40,7 @@
 //
 //M*/
 
-// Trating application for Soft Cascades.
+// Training application for Soft Cascades.
 
 #include <sft/common.hpp>
 #include <iostream>
@@ -114,7 +114,7 @@ int main(int argc, char** argv)
     // 3. Train all octaves
     for (ivector::const_iterator it = cfg.octaves.begin(); it != cfg.octaves.end(); ++it)
     {
-        // a. create rangom feature pool
+        // a. create random feature pool
         int nfeatures  = cfg.poolSize;
         cv::Size model = cfg.model(it);
         std::cout << "Model " << model << std::endl;

modules/softcascade/doc/softcascade_detector.rst

@@ -25,29 +25,37 @@ The sample has been rejected if it fall rejection threshold. So stageless cascad
 .. [BMTG12] Rodrigo Benenson, Markus Mathias, Radu Timofte and Luc Van Gool. Pedestrian detection at 100 frames per second. IEEE CVPR, 2012.
 
 
-SCascade
-----------------
-.. ocv:class:: SCascade
+SoftCascadeDetector
+-------------------
+.. ocv:class:: SoftCascadeDetector
 
 Implementation of soft (stageless) cascaded detector. ::
 
-    class CV_EXPORTS SCascade : public Algorithm
+    class CV_EXPORTS_W SoftCascadeDetector : public Algorithm
     {
     public:
-        SCascade(const float minScale = 0.4f, const float maxScale = 5.f, const int scales = 55, const int rejfactor = 1);
-        virtual ~SCascade();
+
+        enum { NO_REJECT = 1, DOLLAR = 2, /*PASCAL = 4,*/ DEFAULT = NO_REJECT};
+
+        CV_WRAP SoftCascadeDetector(double minScale = 0.4, double maxScale = 5., int scales = 55, int rejCriteria = 1);
+        CV_WRAP virtual ~SoftCascadeDetector();
         cv::AlgorithmInfo* info() const;
-        virtual bool load(const FileNode& fn);
+        CV_WRAP virtual bool load(const FileNode& fileNode);
+        CV_WRAP virtual void read(const FileNode& fileNode);
         virtual void detect(InputArray image, InputArray rois, std::vector<Detection>& objects) const;
-        virtual void detect(InputArray image, InputArray rois, OutputArray rects, OutputArray confs) const;
-    };
+        CV_WRAP virtual void detect(InputArray image, InputArray rois, CV_OUT OutputArray rects, CV_OUT OutputArray confs) const;
 
+    }
 
-SCascade::SCascade
---------------------------
+
+
+SoftCascadeDetector::SoftCascadeDetector
+----------------------------------------
 An empty cascade will be created.
 
-.. ocv:function:: bool SCascade::SCascade(const float minScale = 0.4f, const float maxScale = 5.f, const int scales = 55, const int rejfactor = 1)
+.. ocv:function:: SoftCascadeDetector::SoftCascadeDetector(float minScale = 0.4f, float maxScale = 5.f, int scales = 55, int rejCriteria = 1)
+
+.. ocv:pyfunction:: cv2.SoftCascadeDetector.SoftCascadeDetector(minScale[, maxScale[, scales[, rejCriteria]]]) -> cascade
 
     :param minScale: a minimum scale relative to the original size of the image on which cascade will be applied.
 
@@ -55,35 +63,39 @@ An empty cascade will be created.
 
     :param scales: a number of scales from minScale to maxScale.
 
-    :param rejfactor: used for non maximum suppression.
+    :param rejCriteria: algorithm used for non maximum suppression.
 
 
 
-SCascade::~SCascade
----------------------------
-Destructor for SCascade.
+SoftCascadeDetector::~SoftCascadeDetector
+-----------------------------------------
+Destructor for SoftCascadeDetector.
 
-.. ocv:function:: SCascade::~SCascade()
+.. ocv:function:: SoftCascadeDetector::~SoftCascadeDetector()
 
 
 
-SCascade::load
+SoftCascadeDetector::load
 --------------------------
 Load cascade from FileNode.
 
-.. ocv:function:: bool SCascade::load(const FileNode& fn)
+.. ocv:function:: bool SoftCascadeDetector::load(const FileNode& fileNode)
 
-    :param fn: File node from which the soft cascade are read.
+.. ocv:pyfunction:: cv2.SoftCascadeDetector.load(fileNode)
 
+    :param fileNode: File node from which the soft cascade are read.
 
 
-SCascade::detect
---------------------------
+
+SoftCascadeDetector::detect
+---------------------------
 Apply cascade to an input frame and return the vector of Detection objects.
 
-.. ocv:function:: void SCascade::detect(InputArray image, InputArray rois, std::vector<Detection>& objects) const
+.. ocv:function:: void SoftCascadeDetector::detect(InputArray image, InputArray rois, std::vector<Detection>& objects) const
+
+.. ocv:function:: void SoftCascadeDetector::detect(InputArray image, InputArray rois, OutputArray rects, OutputArray confs) const
 
-.. ocv:function:: void SCascade::detect(InputArray image, InputArray rois, OutputArray rects, OutputArray confs) const
+.. ocv:pyfunction:: cv2.SoftCascadeDetector.detect(image, rois) -> (rects, confs)
 
     :param image: a frame on which detector will be applied.
 
@@ -94,3 +106,40 @@ Apply cascade to an input frame and return the vector of Detection objects.
     :param rects: an output array of bounding rectangles for detected objects.
 
     :param confs: an output array of confidence for detected objects. i-th bounding rectangle corresponds i-th confidence.
+
+
+ChannelFeatureBuilder
+---------------------
+.. ocv:class:: ChannelFeatureBuilder
+
+Public interface for of soft (stageless) cascaded detector. ::
+
+    class CV_EXPORTS_W ChannelFeatureBuilder : public Algorithm
+    {
+    public:
+        virtual ~ChannelFeatureBuilder();
+
+        CV_WRAP_AS(compute) virtual void operator()(InputArray src, CV_OUT OutputArray channels) const = 0;
+
+        CV_WRAP static cv::Ptr<ChannelFeatureBuilder> create();
+    };
+
+
+ChannelFeatureBuilder:~ChannelFeatureBuilder
+--------------------------------------------
+Destructor for ChannelFeatureBuilder.
+
+.. ocv:function:: ChannelFeatureBuilder::~ChannelFeatureBuilder()
+
+
+ChannelFeatureBuilder::operator()
+---------------------------------
+Create channel feature integrals for input image.
+
+.. ocv:function:: void ChannelFeatureBuilder::operator()(InputArray src, OutputArray channels) const
+
+.. ocv:pyfunction:: cv2.ChannelFeatureBuilder.compute(src, channels) -> None
+
+    :param src source frame
+
+    :param channels in OutputArray of computed channels

modules/softcascade/doc/softcascade_training.rst

@@ -1,2 +1,82 @@
 Soft Cascade Training
 =======================
+
+.. highlight:: cpp
+
+Soft Cascade Detector Training
+--------------------------------------------
+
+
+SoftCascadeOctave
+-----------------
+.. ocv:class:: SoftCascadeOctave
+
+Public interface for soft cascade training algorithm
+
+    class CV_EXPORTS SoftCascadeOctave : public Algorithm
+    {
+    public:
+
+        enum {
+            // Direct backward pruning. (Cha Zhang and Paul Viola)
+            DBP = 1,
+            // Multiple instance pruning. (Cha Zhang and Paul Viola)
+            MIP = 2,
+            // Originally proposed by L. Bourdev and J. Brandt
+            HEURISTIC = 4 };
+
+        virtual ~SoftCascadeOctave();
+        static cv::Ptr<SoftCascadeOctave> create(cv::Rect boundingBox, int npositives, int nnegatives, int logScale, int shrinkage);
+
+        virtual bool train(const Dataset* dataset, const FeaturePool* pool, int weaks, int treeDepth) = 0;
+        virtual void setRejectThresholds(OutputArray thresholds) = 0;
+        virtual void write( cv::FileStorage &fs, const FeaturePool* pool, InputArray thresholds) const = 0;
+        virtual void write( CvFileStorage* fs, string name) const = 0;
+
+    };
+
+
+
+SoftCascadeOctave::~SoftCascadeOctave
+---------------------------------------
+Destructor for SoftCascadeOctave.
+
+.. ocv:function:: SoftCascadeOctave::~SoftCascadeOctave()
+
+
+SoftCascadeOctave::train
+------------------------
+
+.. ocv:function:: bool SoftCascadeOctave::train(const Dataset* dataset, const FeaturePool* pool, int weaks, int treeDepth)
+
+    :param dataset an object that allows communicate for training set.
+
+    :param pool an object that presents feature pool.
+
+    :param weaks a number of weak trees should be trained.
+
+    :param treeDepth a depth of resulting weak trees.
+
+
+
+SoftCascadeOctave::setRejectThresholds
+--------------------------------------
+
+.. ocv:function:: void SoftCascadeOctave::setRejectThresholds(OutputArray thresholds)
+
+    :param thresholds an output array of resulted rejection vector. Have same size as number of trained stages.
+
+
+SoftCascadeOctave::write
+------------------------
+
+.. ocv:function:: write SoftCascadeOctave::train(cv::FileStorage &fs, const FeaturePool* pool, InputArray thresholds) const
+.. ocv:function:: write SoftCascadeOctave::train( CvFileStorage* fs, string name) const
+
+    :param fs an output file storage to store trained detector.
+
+    :param pool an object that presents feature pool.
+
+    :param dataset a rejection vector that should be included in detector xml file.
+
+    :param name a name of root node for trained detector.

modules/softcascade/src/soft_cascade_octave.cpp

@@ -11,7 +11,7 @@
 //                For Open Source Computer Vision Library
 //
 // Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
-// Copyright (C) 2008-2012, Willow Garage Inc., all rights reserved.
+// Copyright (C) 2008-2013, Willow Garage Inc., all rights reserved.
 // Third party copyrights are property of their respective owners.
 //
 // Redistribution and use in source and binary forms, with or without modification,
@@ -216,7 +216,7 @@ void BoostedSoftCascadeOctave::setRejectThresholds(cv::OutputArray _thresholds)
 {
     dprintf("set thresholds according to DBP strategy\n");
 
-    // labels desided by classifier
+    // labels decided by classifier
     cv::Mat desisions(responses.cols, responses.rows, responses.type());
     float* dptr = desisions.ptr<float>(0);
 
@@ -423,7 +423,7 @@ void BoostedSoftCascadeOctave::write( cv::FileStorage &fso, const FeaturePool* p
         << "scale" << logScale
         << "weaks" << weak->total
         << "trees" << "[";
-        // should be replased with the H.L. one
+        // should be replaced with the H.L. one
         CvSeqReader reader;
         cvStartReadSeq( weak, &reader);
 
@@ -463,7 +463,7 @@ bool BoostedSoftCascadeOctave::train(const Dataset* dataset, const FeaturePool*
     processPositives(dataset, pool);
     generateNegatives(dataset, pool);
 
-    // 2. only sumple case (all features used)
+    // 2. only simple case (all features used)
     int nfeatures = pool->size();
     cv::Mat varIdx(1, nfeatures, CV_32SC1);
     int* ptr = varIdx.ptr<int>(0);
@@ -471,7 +471,7 @@ bool BoostedSoftCascadeOctave::train(const Dataset* dataset, const FeaturePool*
     for (int x = 0; x < nfeatures; ++x)
         ptr[x] = x;
 
-    // 3. only sumple case (all samples used)
+    // 3. only simple case (all samples used)
     int nsamples = npositives + nnegatives;
     cv::Mat sampleIdx(1, nsamples, CV_32SC1);
     ptr = sampleIdx.ptr<int>(0);
@@ -479,7 +479,7 @@ bool BoostedSoftCascadeOctave::train(const Dataset* dataset, const FeaturePool*
     for (int x = 0; x < nsamples; ++x)
         ptr[x] = x;
 
-    // 4. ICF has an orderable responce.
+    // 4. ICF has an ordered response.
     cv::Mat varType(1, nfeatures + 1, CV_8UC1);
     uchar* uptr = varType.ptr<uchar>(0);
     for (int x = 0; x < nfeatures; ++x)