Viola-Jones人脸检测--AdaptBoost特征选择
Viola-Jones人脸检测算法的伟大之处不不仅仅在于其算法的实时效果，更重要的是其提出了解决目标检测这一类问题的一种通用思路。该算法有两个亮点，一个是积分图技术，一个是Cascade训练模型，一经提出便引起了极大关注，在很多优秀的论文中都能看到他们的身影。如TLD算法中Detector部分，以及BING objectness训练时的两层SVM模型等，很难说这没有受到Viola-Jones算法的影响。下面就来介绍构成Cascade模型的其中的一个基本元素AdaptBoost吧。
AdaptBoost并不是Viola-Jones的原创算法，它是机器学习领域的产物，属于Ensemble Learning中boosting的类别。Ensemble类的学习算法分为bagging和boosting两个类别，都是基于弱分类器构造强分类器的思想，其中bagging的代表算法是RandomForests，boosting的代表算法是AdaptBoost。这里推荐一篇论文，介绍AdaptBoost算法理论的，《A Brief Introduction to Boosting》。
本着分享交流的目的，下面的内容包括对AdaptBoost算法的理论介绍及给出用标准C++实现AdaptBoost的代码。对于不想依赖特定库的伙伴们来说，标准C++的这个版本是个不错的选择。如果有什么不正确的地方，请多多指教。
1.AdaptBoost原理
我们知道对于一个给定窗口大小的图像，其Harr特征的维度是很高的，如果用直接用对训练样本计算出的Harr特征来训练分类器这是不太可行的，我们需要对高维的Harr特征进行选择，选择部分来进行分类器的训练。而AdaptBoost恰好就符合这样的思想，其基本思想是由弱分类器构造强分类器，用弱分类器的联合分类结果作为强分类器的结果。AdaptBoost的弱分类器可以是一个stump，也就是树桩的意思，就是一个弱分类器是一个二分类树。在众多维的Harr特征中进行特征选择的方法是，要求选择一个特征，及选择一个该特征下用于二分类的阈值，如果在该特征和阈值下对训练样本的分类误差最小，就以该特征和其二分类阈值作为一个训练好的弱分类器，算法的具体实现可以参看实现部分的bestStump（）接口。在每一次为弱分类器选择特征完成后，对于用于训练的样本的分布（也就是各样本的权重，初始值一般是相等的，都是1/N，N为样本个数）进行更新，每次的更新是由上一次的弱分类器的分类结果确定的，对于上一次弱分类器判断错误的样本，其权重会增大，判断正确的样本其权重会减小。AdaptBoost与RandomForest的一个区别是，在计算强分类器的结果时，AdaptBoost的弱分类器的权重是不一样的，而RandomForest的弱分类器的权重是相等的。
AdaptBoost算法的伪代码描述如下：
2.标准C++实现
下面的这个接口部分，包含train的接口不包含test的部分，你可以在这个基础上增加test的接口部分。
#ifndef _ADAPTBOOST_H_
#define _ADAPTBOOST_H_
#include <vector>
#include <utility>
#include <cmath>
using namespace std;
/**
* @brief decision stump declaration
*
* @param featureIndex
* @param weightedError achieved weighted error
* @param threshold
* @param margin achieved margin
* @param toggle +1 or -1
*/
struct StumpRule{
int featureIndex;
long double weightedError;
double threshold;
float margin;
int toggle;
};
/**
* @brief what's inside AdaptBoost
*
* @param nPositives number of positive examples
* @param nNegatives number of negative examples
* @param initialPositiveWeight how much weight we give to positives at the outset
* @param ascendingFeatures for each feature, we have (float feature value, int exampleIndex)
*
* @param sampleCount nPositives + nNegatives
* @param inTrain is this a training set or a validation set
* @param exponentialRisk exponential risk for training set
* @param positiveTotalWeight total weight received by positive examples currently
* @param negativeTotalWeight total weight received by negative examples currently
* @param minWeight minimum weight among all weights currently
* @param maxWeight maximum weight among all weights currently
* @param weights weight vector for all examples involved
* @param labels are they positive or negative examples
* @param featureCount how many features are there
* @param committee what's the learned committee
*/
class AdaptBoost{
private:
int nPositives;
int nNegatives;
long double initialPositiveWeight;
vector< vector<pair<float, int>> > ascendingFeatures;
int sampleCount;
int featureCount;
long double positiveTotalWeight;
long double negativeTotalWeight;
long double minWeight;
long double maxWeight;
long double exponentialRisk;
vector<double> weights;
vector<int> labels;
vector<StumpRule> committee;
/**
* @brief prevent copy and assignment
*/
AdaptBoost(const AdaptBoost&);
AdaptBoost operator=(const AdaptBoost&);
protected:
/**
* @brief return for an element pointed by iterator and featureIndex its exampleIndex
*/
int getTrainingExampleIndex(int featureIndex, int iterator);
/**
* @brief return for an element pointed by iterator and featureIndex its example value
*/
float getTrainingExampleFeature(int featureIndex, int iterator);
/**
* @brief sort each featrue from different samples
*/
void sortFeatures(
vector< vector<pair<float, int>> >& features
);
/**
* @brief best stump given a feature
*/
void decisionStump(
int featureIndex
, StumpRule & best
);
/**
* @brief best stump among all features
*/
StumpRule bestStump();
public:
/**
* @brief constructor
* @param nPositives number of positives for training examples
* @param nNegatives number of negatives for training examples
* @param initialPositiveWeight initial weight of positives
* @param data for training examples, positves front and negatives back
*/
AdaptBoost(
int nPositives
, int nNegatives
, long double initialPositiveWeight
, const vector< vector<float> >& data
);
/**
* @brief destructor
*/
~AdaptBoost();
/**
* @brief perform one round of adaboost
*/
void oneRoundOfAdaboostTraining();
/**
* @brief get committee adaptboost trained
*/
vector<StumpRule> getCommittee() {
return committee;
}
/**
* @brief get committee size
*/
int getCommitteeSize() {
return committee.size();
}
/**
* @brief given the number of weak classifiers train for a committee
* @param numOfWeakClassifier for number of weak classifiers of adapt boost
*/
void adaptBoostTraining(int numOfWeakClassifier);
/**
* @brief evaluate how the committee fares on a training dataset
*
* @param tweak for predictLableOfTrainingExamples
* @return falsePositive
* @return detectionRate
* @vector<int> return a blackList,if element of balckList is 0, then it means that
* this sample could be used again otherwise it means not usable
*/
vector<int> calcEmpiricalErrorInAdaBoostTraining(
float tweak
, float & falsePositive
, float & detectionRate
);
/**
* @brief given a tweak and a committe, what prediction do you make as to the training examples
*
* @param thresholdTweak tweak
* @return prediction
* @param onlyMostRecent use all the committee or its most recent member (a weak learner)
*/
void predictLabelOfTrainingExamples(
float tweakThreshold
, vector<int> & prediction
, bool onlyMostRecent=false
);
};
#endif
#include <cassert>
#include <algorithm>
#include <iostream>
#include <iomanip>
#include "VJAdaptBoost.h"
using namespace std;
#define VERBOSE true
//fail and messaging
static void fail(const char* message){
cerr << "Error:" << message << endl;
exit(EXIT_FAILURE);
}
//order definition for this type of pairs
//compare only the feature values
static bool myPairOrder(
const pair<float, int>& one
, const pair<float, int>& other
){
return one.first < other.first;
}
//why is one stump better than the other
static bool myStumpOrder(
const StumpRule & one
, const StumpRule & other
){
if(one.weightedError < other.weightedError)
return true;
if(one.weightedError == other.weightedError && one.margin > other.margin)
return true;
return false;
}
int AdaptBoost::getTrainingExampleIndex(int featureIndex, int iterator){
assert(ascendingFeatures.size() > 0 && ascendingFeatures[0].size() >0);
return ascendingFeatures[featureIndex][iterator].second;
}
float AdaptBoost::getTrainingExampleFeature(int featureIndex, int iterator){
assert(ascendingFeatures.size() > 0 && ascendingFeatures[0].size() >0);
if(_isnan(ascendingFeatures[featureIndex][iterator].first)){
cerr<<"ERROR: nan feature "<<featureIndex<<" detected for example "<<getTrainingExampleIndex(featureIndex, iterator)<<endl;
exit(EXIT_FAILURE);
}
return ascendingFeatures[featureIndex][iterator].first;
}
//constructor
AdaptBoost::AdaptBoost(
int positives
, int negatives
, long double positiveWeight
, const vector< vector<float> >& data) {
assert(positives > 0 && negatives > 0);
assert(positiveWeight > 0 && positiveWeight < 1);
assert(data.size() > 0 && data[0].size() > 0 );
assert(data.size() == (positives + negatives));
//add number of data info to features
vector< vector<pair<float, int>> > features(data.size(), vector<pair<float, int>>(data[0].size(), pair<float, int>(0,0)));
for(int i=0; i<features.size(); i++) {
for(int j=0; j<features[0].size(); j++) {
features[i][j] = pair<float, int>(data[i][j], i);
}
}
//initialize the class attributes for the training set
nPositives = positives;
nNegatives = negatives;
initialPositiveWeight = positiveWeight;
sortFeatures(features);//initialize ascendingFeatures
sampleCount = positives + negatives;
featureCount = ascendingFeatures.size();
positiveTotalWeight = positiveWeight;
negativeTotalWeight = 1 - positiveWeight;
long double posAverageWeight = positiveTotalWeight/(long double)nPositives;
long double negAverageWeight = negativeTotalWeight/(long double)nNegatives;
maxWeight = max(posAverageWeight, negAverageWeight);
minWeight = min(posAverageWeight, negAverageWeight);
exponentialRisk = 1;
//set weights for each example
for(int exampleIndex = 0; exampleIndex < sampleCount; exampleIndex++){
weights.push_back(exampleIndex < nPositives ? posAverageWeight : negAverageWeight);
labels.push_back(exampleIndex < nPositives ? 1 : -1);
}
}
//destructor
AdaptBoost::~AdaptBoost() {
}
//adaptBoost interface for training
void AdaptBoost::adaptBoostTraining(int numOfWeakClassifier) {
assert(numOfWeakClassifier > 0);
for(int i=0; i<numOfWeakClassifier; i++) {
oneRoundOfAdaboostTraining();
}
}
//validation procedure using training examples
vector<int> AdaptBoost::calcEmpiricalErrorInAdaBoostTraining(
float tweak
, float & falsePositive
, float & detectionRate
){
vector<int> blackList;
blackList.resize(nPositives, 0);
blackList.resize(nPositives+nNegatives, 1);
int nFalsePositive = 0;
int nFalseNegative = 0;
//initially let all be positive
vector<int> prediction;
prediction.resize(sampleCount,0);
predictLabelOfTrainingExamples(tweak, prediction, false);
//evaluate prediction errors
vector<int> agree(sampleCount);
for(int i=0; i<sampleCount; i++) {
agree[i] = labels[i]*prediction[i];
}
for(int exampleIndex = 0; exampleIndex < sampleCount; exampleIndex++){
if(agree[exampleIndex] < 0) {
if(exampleIndex < nPositives){
nFalseNegative += 1;
blackList[exampleIndex] = 1;
}else{
nFalsePositive += 1;
blackList[exampleIndex] = 0;
}
}
}
//set the returned values
falsePositive = nFalsePositive/(float)nNegatives;
detectionRate = 1 - nFalseNegative/(float)nPositives;
return blackList;
}
//given a tweak and a committe, what prediction does it make as to the training examples
void AdaptBoost::predictLabelOfTrainingExamples(
float tweakThreshold
, vector<int> & prediction
, bool onlyMostRecent
){
int committeeSize = committee.size();
//no need to weigh a single member's decision
onlyMostRecent = committeeSize == 1 ? true : onlyMostRecent;
int start = onlyMostRecent ? committeeSize - 1 : 0;
//double to be more precise
vector<vector<double>> memberVerdict;
for(int i=0; i<committeeSize; i++) {//initialize memberVerdict
vector<double> row(sampleCount);
memberVerdict.push_back(row);
}
vector<double> memberWeight(committeeSize);
//members, go ahead
for(int member = start; member < committeeSize; member++){
//sanity check
if(committee[member].weightedError == 0 && member != 0)
fail("Boosting Error Occured!");
//0.5 does not count here
//if member's weightedError is zero, member weight is nan, but it won't be used anyway
memberWeight[member] = log(1./committee[member].weightedError -1);
int feature = committee[member].featureIndex;
#pragma omp parallel for schedule(static)
for(int iterator = 0; iterator < sampleCount; iterator++){
int exampleIndex = getTrainingExampleIndex(feature, iterator);
memberVerdict[member][exampleIndex] = (getTrainingExampleFeature(feature, iterator) >
committee[member].threshold ? 1 : -1)*committee[member].toggle + tweakThreshold;
}
}
//joint session
if(!onlyMostRecent){
vector<double> finalVerdict(sampleCount);
for(int i=0; i<sampleCount; i++) {
double predict = 0;
for(int j=0; j<committeeSize; j++) {
predict += (memberWeight[j] * memberVerdict[j][i]);
}
finalVerdict[i] = predict;
}
for(int exampleIndex = 0; exampleIndex < sampleCount; exampleIndex++)
prediction[exampleIndex] = finalVerdict[exampleIndex] > 0 ? 1 : -1;
}else{
for(int exampleIndex = 0; exampleIndex < sampleCount; exampleIndex++)
prediction[exampleIndex] = memberVerdict[start][exampleIndex] > 0 ? 1 : -1;
}
}
void AdaptBoost::oneRoundOfAdaboostTraining(){
//try to be friendly here
static int trainPhase = 0;
if(VERBOSE && trainPhase == 0){
cout << "\n#############################ADABOOST MESSAGE EXPLAINED####################################################\n\n";
cout << "INFO: Adaboost starts. Exponential Risk is expected to go down steadily and strictly," << endl;
cout << "INFO: and Exponential Risk should bound the (weighted) Empirical Error from above." << endl;
cout << "INFO: Train Phase is the current boosting iteration." << endl;
cout << "INFO: Best Feature is the most discriminative feature selected by decision stump at this iteration." << endl;
cout << "INFO: Threshold and Toggle are two parameters that define a real valued decision stump.\n" << endl;
}
trainPhase++;
//get and store the rule
StumpRule rule = bestStump();
committee.push_back(rule);
//how it fares
vector<int> prediction(sampleCount);
predictLabelOfTrainingExamples(
0
, prediction
, /*onlyMostRecent*/ true);
vector<bool> agree(sampleCount);
for(int i=0; i<sampleCount; i++) {
if(prediction[i] == labels[i]) {
agree[i] = true;
}else {
agree[i] = false;
}
}
//update weights
vector<double> weightUpdate;
weightUpdate.resize(sampleCount,1);
bool errorFlag = false;
for(int exampleIndex = 0; exampleIndex < sampleCount; exampleIndex++){
//more weight for a difficult example
if(!agree[exampleIndex]){
weightUpdate[exampleIndex] = 1/rule.weightedError - 1;
errorFlag = true;
}
}
//update weights only if there is an error
if(errorFlag){
double weightSum = 0;
for(int i=0; i<sampleCount; i++) {
weights[i] *= weightUpdate[i];
weightSum += weights[i];
}
for(int i=0; i<sampleCount; i++) {
weights[i] /= weightSum;
}
double posTotalWeight = 0;
for(int i=0; i<nPositives; i++) {
posTotalWeight += weights[i];
}
positiveTotalWeight = posTotalWeight;
negativeTotalWeight = 1-positiveTotalWeight;
double min,max;
min = max = weights[0];
for(int i=0; i<sampleCount; i++) {
if(weights[i] < min) {
min = weights[i];
}else if(weights[i] > max) {
max = weights[i];
}
}
minWeight = min;
maxWeight = max;
}
//exponentialRisk can be zero at the first boosting
exponentialRisk *= 2*sqrt((1-rule.weightedError)*rule.weightedError);
//print some statistics
if(VERBOSE){
float tweak = 0;
float falsePositive = 0;
float detectionRate = 0;
calcEmpiricalErrorInAdaBoostTraining(tweak, falsePositive, detectionRate);
float empError = static_cast<float>(falsePositive*(1-initialPositiveWeight)+initialPositiveWeight*(1-detectionRate));
cout << "Training Performance Explanation (before threshold tweaking): falsePositive " << falsePositive
<< " detectionRate " << detectionRate << endl;
cout <<"###########################################################################################################\n";
cout << "\nTrain Phase " << trainPhase << endl << endl;
// whatFeature(rule.featureIndex);
cout << "\tExponential Risk " << setw(12) << exponentialRisk << setw(19) << "Weighted Error "
<< setw(11) << rule.weightedError << setw(14) << "Threshold " << setw(10) << rule.threshold
<< setw(13) <<"Toggle " << setw(12) << rule.toggle << endl;
cout << "\tPositive Weight" << setw(14) << positiveTotalWeight << setw(14) << "MinWeight "
<< setw(16) << minWeight << setw(14) << "MaxWeight " << setw(10) << maxWeight << setw(22)
<< "Empirical Error " << setw(10) << empError << endl << endl;
}
}
//get a feature from features and put them in ascending order
//and record at the same time the permuted example order
void AdaptBoost::sortFeatures(vector< vector<pair<float, int>> >& features) {
assert(features.size()!=0 && features[0].size() !=0 );
for(unsigned int i=0; i<features[0].size(); i++) {
vector<pair<float, int>> temp = vector<pair<float, int>>();
for(unsigned int j=0; j<features.size(); j++) {
temp.push_back(features[j][i]);
}
//sort
sort(temp.begin(), temp.end(), myPairOrder);
ascendingFeatures.push_back(temp);
}
}
//base learner is a stump, a decision tree of depth 1
//decisionStump has to look at feature and return rule
void AdaptBoost::decisionStump(
int featureIndex
, StumpRule & best
){
//a stump is determined by threshold and toggle, the other two attributes measures its performance
//initialize with some crazy values
best.featureIndex = featureIndex;
best.weightedError = 2;
best.threshold = getTrainingExampleFeature(featureIndex, 0) - 1;
best.margin = -1;
best.toggle = 0;
StumpRule current = best;
//error_p and error_n allow to set the best toggle
long double error_p, error_n;
//initialize: r denotes right hand side and l left hand side
//convention: in TrainExamples nPositives positive samples are followed by negatives samples
long double rPositiveWeight = positiveTotalWeight;
long double rNegativeWeight = negativeTotalWeight;
//yes, nothing to the left of the sample with the smallest feature
long double lPositiveWeight = 0;
long double lNegativeWeight = 0;
//go through all these observations one after another
int iterator = -1;
//to build a decision stump, you need a toggle and an admissible threshold
//which doesn't coincide with any of the observations
while(true){
//We've got a threshold. So determine the best toggle based on two types of error
//toggle = 1, positive prediction if and only if the observed feature > the threshold
//toggle = -1, positive prediction if and only if the observed feature < the threshold
//error_p denotes the error introduced by toggle = 1, error_n the error by toggle = -1
error_p = rNegativeWeight + lPositiveWeight;
error_n = rPositiveWeight + lNegativeWeight;
current.toggle = error_p < error_n ? 1 : -1;
//sometimes shit happens, prevent error from being negative
long double smallerError = min(error_p, error_n);
//this prevents some spurious nonzero: for currentError must be at least equal to minWeight
current.weightedError = smallerError < minWeight * 0.9 ? 0 : smallerError;
//update if necessary
if(myStumpOrder(current, best))
best = current;
//move on
iterator++;
//we don't actually need to look at the sample with the largest feature
//because its rule is exactly equivalent to those produced
//by the sample with the smallest feature on training observations
//but it won't do any harm anyway
if(iterator == sampleCount)
break;
//handle duplicates, update lr weights and find a new threshold
while(true){
//take this guy's attributes
int exampleIndex = getTrainingExampleIndex(featureIndex, iterator);
int label = labels[exampleIndex];
long double weight = weights[exampleIndex];
//update weights
if(label < 0){
lNegativeWeight += weight;
rNegativeWeight -= weight;
}else{
lPositiveWeight += weight;
rPositiveWeight -= weight;
}
//if a new threshold can be found, break
//two cases are possible: either it is the last observation
if(iterator == sampleCount - 1)
break;
//or no duplicate. If there is a duplicate, repeat
if(getTrainingExampleFeature(featureIndex, iterator) != getTrainingExampleFeature(featureIndex, iterator + 1)){
double test = ((double)getTrainingExampleFeature(featureIndex, iterator)
+ (double)getTrainingExampleFeature(featureIndex, iterator + 1))/2;
//well that's a bit frustrating: I want to keep float because of memory constraint, but apparently
//features are so close, sometimes, numerical precision arises as an unexpected problem, so I decide
//to use a double threshold so as to separate float features
if(getTrainingExampleFeature(featureIndex, iterator) < test && test < getTrainingExampleFeature(featureIndex, iterator + 1))
break;
else{
#pragma omp critical
{
cout << "ERROR: numerical precision breached: problem feature values "
<< getTrainingExampleFeature(featureIndex, iterator)
<< " : " << getTrainingExampleFeature(featureIndex, iterator+1)
<< ". Problem feature " << featureIndex << " and problem example "
<< getTrainingExampleIndex(featureIndex, iterator) << " : "
<< getTrainingExampleIndex(featureIndex, iterator+1) << endl;
}
fail("fail to find a suitable threshold.");
}
}
iterator++;
}
//update threshold
if(iterator < sampleCount - 1){
current.threshold = ((double)getTrainingExampleFeature(featureIndex, iterator)
+ (double)getTrainingExampleFeature(featureIndex, iterator + 1))/2;
current.margin = getTrainingExampleFeature(featureIndex, iterator + 1) - getTrainingExampleFeature(featureIndex, iterator);
}else{
//slightly to the right of the biggest observation
current.threshold = getTrainingExampleFeature(featureIndex, iterator) + 1;
current.margin = 0;
}
}
}
//implement the feature selection's outer loop
//return the most discriminative feature and its rule
StumpRule AdaptBoost::bestStump(
){
vector<StumpRule> candidates;
candidates.resize(featureCount);
#pragma omp parallel for schedule(static)
for(int featureIndex = 0; featureIndex < featureCount; featureIndex++)
decisionStump(featureIndex, candidates[featureIndex]);
//loop over all the features
//the best rule has the smallest weighted error and the largest margin
StumpRule best = candidates[0];
for(int featureIndex = 1; featureIndex < featureCount; featureIndex++){
if(myStumpOrder(candidates[featureIndex], best))
best = candidates[featureIndex];
}
//if shit happens, tell me
if( best.weightedError >= 0.5 )
fail("Decision Stump failed: base error >= 0.5");
//return
return best;
}
reference：
Yi-Qing Wang, An Analysis of the Viola-Jones Face Detection Algorithm, IPOL.
本文出自 “Remys” 博客，谢绝转载！