// 第一種方法提取整幅圖像的ULBP特徵
// 計算ULBP索引表
bool ulbpIndex(vector<int> &uniform_lbp)
{uniform_lbp.clear();
for (int i = 0; i < 256; ++i)
{
int data = i;
int dataTmp1 = data, dataTmp2 = data;
int jump = 0; int tmp = -1;
for (int k = 0; k < 8; ++k)
{
dataTmp1 = data >> 1;
dataTmp2 = dataTmp1 << 1;
int t = data ^ dataTmp2;
data = dataTmp1;
assert(t == 0 || t == 1);
if (tmp == -1 && k == 0)
{
tmp = t;
continue;
}
if (tmp != t)
{
jump++;
tmp = t;
}
}
if (jump < 3)
{
uniform_lbp.push_back(i);
}
}
if (uniform_lbp.size() != 58)
{
return false;
}
return true;
}
// ULBP特徵提取
template <typename _Tp> static
inline void ulbp_(InputArray _src, OutputArray _dst, int radius, int neighbors, vector<int> m_uniform)
{
if (neighbors != 8 || m_uniform.size() != 58)
{
cout << "neighbor must be 8! and uniform size be 58!\n";
system("pause");
exit(-1);
}
//get matrices
Mat src = _src.getMat();
// allocate memory for result
_dst.create(src.rows-2*radius, src.cols-2*radius, CV_32SC1);
Mat dst = _dst.getMat();
// zero
dst.setTo(0);
for(int n=0; n<neighbors; n++)
{
// sample points
float x = static_cast<float>(-radius) * sin(2.0*CV_PI*n/static_cast<float>(neighbors));
float y = static_cast<float>(radius) * cos(2.0*CV_PI*n/static_cast<float>(neighbors));
// relative indices
int fx = static_cast<int>(floor(x));
int fy = static_cast<int>(floor(y));
int cx = static_cast<int>(ceil(x));
int cy = static_cast<int>(ceil(y));
// fractional part
float ty = y - fy;
float tx = x - fx;
// set interpolation weights
float w1 = (1 - tx) * (1 - ty);
float w2 = tx * (1 - ty);
float w3 = (1 - tx) * ty;
float w4 = tx * ty;
// iterate through your data
for(int i=radius; i < src.rows-radius;i++) {
for(int j=radius;j < src.cols-radius;j++) {
// calculate interpolated value
float t = w1*src.at<_Tp>(i+fy,j+fx) + w2*src.at<_Tp>(i+fy,j+cx) + w3*src.at<_Tp>(i+cy,j+fx) + w4*src.at<_Tp>(i+cy,j+cx);
// floating point precision, so check some machine-dependent epsilon
dst.at<int>(i-radius,j-radius) += ((t > src.at<_Tp>(i,j)) ||
(std::abs(t-src.at<_Tp>(i,j)) < std::numeric_limits<float>::epsilon()))
<< n;
}
}
}
for (int i = 0; i < dst.rows; ++i)
{
for (int j = 0; j < dst.cols; ++j)
{
int data = dst.at<int>(i, j);
vector<int>::iterator iter = find(m_uniform.begin(), m_uniform.end(), data);
if (iter == m_uniform.end())
{
dst.at<int>(i, j) = 0;
}
else
{
int new_data = iter - m_uniform.begin() ;
dst.at<int>(i, j) = new_data + 1;
}
}
}
}
}
// 第二種方法提取關鍵點附近的ULBP特徵
// ULBP特徵映射表
int mapping[256] =
{
0, 1, 2, 3, 4, 58, 5, 6, 7, 58, 58, 58, 8, 58, 9, 10,
11, 58, 58, 58, 58, 58, 58, 58, 12, 58, 58, 58, 13, 58, 14, 15,
16, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58,
17, 58, 58, 58, 58, 58, 58, 58, 18, 58, 58, 58, 19, 58, 20, 21,
22, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58,
58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58,
23, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58,
24, 58, 58, 58, 58, 58, 58, 58, 25, 58, 58, 58, 26, 58, 27, 28,
29, 30, 58, 31, 58, 58, 58, 32, 58, 58, 58, 58, 58, 58, 58, 33,
58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 34,
58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58,
58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 35,
36, 37, 58, 38, 58, 58, 58, 39, 58, 58, 58, 58, 58, 58, 58, 40,
58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 58, 41,
42, 43, 58, 44, 58, 58, 58, 45, 58, 58, 58, 58, 58, 58, 58, 46,
47, 48, 58, 49, 58, 58, 58, 50, 51, 52, 58, 53, 54, 55, 56, 57
};
//Compute an single lbp code from a pixel
int LBPFeatureExtractor::lbpCode (unsigned char seq[9])
{
bool lab[8] = {false};
int base = seq[0];
int result = 0, one = 1, final;
int i;
for (i = 0; i < 8; i++) {
if (base >= seq[i+1])
lab[i] = 1;
else
lab[i] = 0;
}
for (i = 0; i < 8; i++, one = one << 1)
result += lab[i]*one;
if (result >= 256)
{
printf("LBP編碼錯誤\n");
system("pause");
exit(-1);
}
return mapping[result];
}
//Compute lbp histgoram from an image
int* lbpHist(Mat &img, int* lbpHist)
{
unsigned char locP[9];
for(int i=0; i<59; i++)
lbpHist[i] = 0;
for(int i = 1; i<img.rows-1; i++)
for(int j = 1; j< img.cols-1; j++)
{
locP[0] = img.at<unsigned char>(i,j);
locP[1] = img.at<unsigned char>(i-1,j);
locP[2] = img.at<unsigned char>(i-1,j-1);
locP[3] = img.at<unsigned char>(i,j-1);
locP[4] = img.at<unsigned char>(i+1,j-1);
locP[5] = img.at<unsigned char>(i+1,j);
locP[6] = img.at<unsigned char>(i+1,j+1);
locP[7] = img.at<unsigned char>(i,j+1);
locP[8] = img.at<unsigned char>(i-1,j+1);
lbpHist[lbpCode(locP)]++;
}
return lbpHist;
}
int* extractAt(const Mat &inputImage, const vector< pair<double, double> >points, int *outputFeature){
Mat processedImage;
Mat resizeImage;
//convert image into graylevel
// 將彩色圖像轉換爲灰度圖像
if (inputImage.channels() == 3) {
cvtColor(inputImage, processedImage, CV_BGR2GRAY );
} else {
processedImage = inputImage;
}
equalizeHist(processedImage, processedImage);
// 對於每個尺度下的圖像提取LBP特徵
int count = 0;
for(int s = 0; s<this->scale.size(); s++)
{
int currentScale = this->scale[s];
//scale face into proper size
// 將關鍵點映射到當前尺度
vector< pair<double, double> > newPoints;
for(int i=0; i<points.size(); i++)
{
pair<double,double> point;
point.first = points[i].first*double(currentScale)/inputImage.cols;
point.second = points[i].second*double(currentScale)/inputImage.rows;
newPoints.push_back(point);
}
//printf("%d %d %d\n", currentScale, inputImage.cols, inputImage.rows);
// 直方圖均衡化後縮放圖像，將直方圖均衡化放到縮放尺度之外，沒有必要每次都做
// equalizeHist(processedImage, processedImage);
resize(processedImage, resizeImage, Size(currentScale, currentScale));
//compute center and extract lbp feature
Mat patch;
int hist[59];
// 對於每個關鍵點提取，patchsize=10，numCellX=4,numCellY=4
for(int i=0; i<newPoints.size(); i++)
{
for(int j=0; j<numCellX; j++)
for(int k=0; k<numCellY; k++)
{
double centerX = int(newPoints[i].first - patchSize*(numCellX/2) + (numCellX%2 == 0)*patchSize*0.5 + patchSize*j);
double centerY = int(newPoints[i].second - patchSize*(numCellY/2) + (numCellY%2 == 0)*patchSize*0.5 + patchSize*k);
//circle(resizeImage, Point(centerX,centerY), 5, Scalar(255,0,0), CV_FILLED);
getRectSubPix(resizeImage, Size(patchSize+2, patchSize+2), Point2f(centerX,centerY), patch);
lbpHist(patch, hist);
for(int l=0; l<59; l++)
outputFeature[count++] = hist[l];
}
}
//imwrite(to_string(currentScale)+"out.jpg", resizeImage);
resizeImage.release();
}
processedImage.release();
return NULL;
}