haarlearner.h

来自「MultiBoost 是c++实现的多类adaboost酸法。与传统的adabo」· C头文件 代码 · 共 177 行

/*
* This file is part of MultiBoost, a multi-class 
* AdaBoost learner/classifier
*
* Copyright (C) 2005-2006 Norman Casagrande
* For informations write to [e-mail]
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
*
*/

/**
* \file HaarLearner.h A generic interface for Haar-Like features.
*/

#ifndef __HAAR_LEARNER_H
#define __HAAR_LEARNER_H

#include <string>
#include <vector>
#include <iosfwd>

#include "Utils/Utils.h" // for rect


using namespace std;

//////////////////////////////////////////////////////////////////////////////////////////////
//////////////////////////////////////////////////////////////////////////////////////////////

// forward declarations to save some includes
namespace nor_utils {
   class StreamTokenizer;
   class Args;
}

namespace MultiBoost {

// forward declarations to save some includes
class HaarFeature;
class HaarData;

/**
* A generic interface to the learner based on Haar-like features.
* These features, first described in an AdaBoost environment by Viola and
* Jones in their 2001 paper (Robust Real-time Object Detection) are
* powerful way to treat with 1 or 2-dimensional data. This class
* it is \b not a learner itself, but must be extended among with another
* class that extend BaseLearner. For instance, the class
* HaarSingleStumpLearner extends SingleStumpLearner \b and HaarLearner.
* 
* This class contains the generic operation for all the Haar-based
* learners.
* @date 16/12/2005
*/
class HaarLearner
{
public:

   /**
   * The constructor. Initializes the attributes belonging to all Haar-like based
   * learners.
   * @date 16/12/2005
   */
   HaarLearner() : 
      _pSelectedFeature(NULL), _samplingType(ST_NO_SAMPLING), _samplingVal(0) {}

   /**
   * The destructor. Must be declared (virtual) for the proper destruction of 
   * the object.
   */
   virtual ~HaarLearner() {}

   /**
   * Declare Haar-learner-specific arguments.
   * @remark It must be explicitly called by the derived class.
   * @param args The Args class reference which can be used to declare
   * additional arguments.
   * @date 16/12/2005
   */
   void declareArguments(nor_utils::Args& args);

   /**
   * Set the arguments of the algorithm using the standard interface
   * of the arguments. Call this to set the arguments asked by the user.
   * @remark It must be explicitly called by the derived class.
   * @param args The arguments defined by the user in the command line.
   * @date 16/12/2005
   */
   void initOptions(nor_utils::Args& args);

   /**
   * Save the current object information needed for classification.
   * @param outputStream The stream where the data will be saved
   * @param numTabs The number of tabs before the tag. Useful for indentation
   * @remark It must be explicitly called by the derived class.
   * @see StumpLearner::save()
   * @date 16/12/2005
   */
   virtual void save(ofstream& outputStream, int numTabs = 0);

   /**
   * Load the xml file that contains the serialized information
   * needed for the classification and that belongs to this class.
   * @remark It must be explicitly called by the derived class.
   * @param st The stream tokenizer that returns tags and values as tokens
   * @see save()
   * @date 16/12/2005
   */
   virtual void load(nor_utils::StreamTokenizer& st);

protected:

   /**
   * The types of sampling techniques in the case not all the configurations are
   * evaluated.
   * @date 16/12/2005
   */
   enum eSamplingType
   {
      ST_NO_SAMPLING, //!< Default.
      ST_NUM, //!<  A limited number per feature type, per iteration.
      ST_TIME //!< A limited time per feature type, per iteration.
   };

   nor_utils::Rect _selectedConfig; //!< The selected configuration of the feature. 

   /**
   * The selected feature type. This is a copy of the pointer to the loaded features types,
   * and it is only used to access the Haar-computation function that belongs to the type.
   * For instance, if during the current iteration a 2 blocs vertical feature is selected
   * the pointer to this type of feature is copied here. During classification, this
   * object is used to obtain the haar value computed with the Haar-like feature.
   * @see HaarData::_loadedFeatures
   * @see HaarFeature::getValue
   * @date 17/12/2005
   */
   HaarFeature*    _pSelectedFeature; //!< The selected feature type. A pointer to the  

   /**
   * The type of sampling.
   * @see eSamplingType
   * @date 17/12/2005
   */
   eSamplingType _samplingType; 

   /**
   * The value associated the the sampling type.
   *
   * If \a _samplingType is ST_NUM, this variable holds the number of sampling
   * configurations to be computed per feature, per boosting iteration.
   *
   * If \a _samplingType is ST_TIME, this variable holds the time limit for
   * evaluating the configurations. This time is per configuration, per boosting
   * iteration.
   * @date 17/12/2005
   */
   int           _samplingVal;

};

} // end of namespace MultiBoost

#endif // __HAAR_LEARNER_H