基于MFC开发的指纹识别系统.
MFC-FingerPrint
基于MFC开发的指纹识别系统.
效果图如下:
在第12步特征入库中，会对当前指纹的mdl数据与databases中所有的mdl进行对比,然后返回识别结果.
一.载入图像
在点击之后,选择需要识别的图片. 图片路径为beginfilename.
void CFingerDlg::OnBnClickedOk1()
{
CFileDialog dlgFile(TRUE, NULL, NULL, OFN_HIDEREADONLY, _T("Describe Files (*.bmp)|*.bmp|All Files (*.*)|*.*||"), NULL);
if (dlgFile.DoModal())
{
strFile = dlgFile.GetPathName();
}
char *info = "";
USES_CONVERSION;
char * beginfilename = T2A(strFile);
Step1_LoadBmpImage(beginfilename,info);
ShowImageInCtrl(m_picCtrl1, beginfilename);
// TODO: 在此添加控件通知处理程序代码
}
Step1_LoadBmpImage中就是获取初始图片的RGB信息.
int Step1_LoadBmpImage(char *beginfilename,char* info) {
char *filename = beginfilename;
CopyFile(ToWideChar(filename), ToWideChar(STEP_TXT_1), false);
int iWidth, iHeight, iDepth;
int flag = ReadBMPImgFilePara(filename, iWidth, iHeight, iDepth);
if (flag != 0) {
//sprintf(info,"图像加载失败");
::MessageBox(NULL, _T("图像加载失败"), _T("error"), MB_OK);
return -1;
}
unsigned char *data = new unsigned char[iWidth*iHeight];
flag = ReadBMPImgFileData(filename, data);
if (flag != 0) {
//sprintf(info, "图像数据读取失败");
::MessageBox(NULL, _T("图像数据读取失败"), _T("error"), MB_OK);
delete[] data;
return -2;
}
flag = SaveDataToTextFile(STEP_TXT_1, data, iWidth, iHeight);
if (flag != 0) {
//sprintf(info,"数据保存失败");
::MessageBox(NULL, _T("数据保存失败"), _T("error"), MB_OK);
delete[] data;
return -3;
}
//sprintf(info, "源图[%s],宽度[%d]，高度[%d]，深度[%d b]",filename,iWidth,iHeight,iDepth);
delete[] data;
return 0;
}
其中具体的Txt与BMP转换参考具体源代码.
二.中值滤波
中值滤波实现算法,ucImg由第一步得来.
int MidFilter(unsigned char *ucImg, unsigned char *ucDstImg, int iWidth, int iHeight) {
memset(ucDstImg, 0, iWidth*iHeight);
unsigned char *pUp, *pDown, *pImg;
unsigned char x[9];
for (int i = 1; i < iHeight - 1; i++) {
pUp = ucImg + (i - 1)*iWidth;
pImg = ucImg + i * iWidth;
pDown = ucImg + (i + 1)*iWidth;
int j;
for (j = 1; j < iWidth - 1; j++) {
pUp++;
pImg++;
pDown++;
x[0] = *(pUp - 1);
x[1] = *(pImg - 1);
x[2] = *(pDown - 1);
x[3] = *pUp;
x[4] = *pImg;
x[5] = *pDown;
x[6] = *(pUp + 1);
x[7] = *(pImg + 1);
x[8] = *(pDown + 1);
Sort(x, 9);
*(ucDstImg + i * iWidth + j) = x[4];
}
}
int j;
pDown = ucImg + iWidth;
for (j = 1; j < iWidth - 1; j++) {
x[0] = *(ucImg + j - 1);
x[1] = *(ucImg + j);
x[2] = *(ucImg + j + 1);
x[3] = *(pDown + j - 1);
x[4] = *(pDown + j);
x[5] = *(pDown + j + 1);
Sort(x, 6);
*(ucDstImg + j) = x[3];
}
pUp = ucImg + iWidth * (iHeight - 2);
pDown = ucImg + iWidth * (iHeight - 1);
for (j = 1; j < iWidth - 1; j++) {
x[0] = *(pDown + j - 1);
x[1] = *(pDown + j);
x[2] = *(pDown + j + 1);
x[3] = *(pUp + j - 1);
x[4] = *(pUp + j);
x[5] = *(pUp + j + 1);
Sort(x, 6);
*(ucDstImg + iWidth * (iHeight - 1) + j) = x[3];
}
x[0] = *(ucImg);
x[1] = *(ucImg + 1);
x[2] = *(ucImg + iWidth);
x[3] = *(ucImg + iWidth + 1);
Sort(x, 4);
*(ucDstImg) = x[2];
x[0] = *(ucImg + iWidth - 1);
x[1] = *(ucImg + iWidth - 2);
x[2] = *(ucImg + 2 * iWidth - 1);
x[3] = *(ucImg + 2 * iWidth - 2);
Sort(x, 4);
*(ucDstImg + iWidth - 1) = x[2];
x[0] = *(ucImg + iWidth * (iHeight - 1));
x[1] = *(ucImg + iWidth * (iHeight - 2));
x[2] = *(ucImg + iWidth * (iHeight - 1) + 1);
x[3] = *(ucImg + iWidth * (iHeight - 2) + 1);
Sort(x, 4);
*(ucDstImg + (iHeight - 1)*iWidth) = x[2];
x[0] = *(ucImg + iWidth * (iHeight - 0) - 1);
x[1] = *(ucImg + iWidth * (iHeight - 1) - 1);
x[2] = *(ucImg + iWidth * (iHeight - 0) - 2);
x[3] = *(ucImg + iWidth * (iHeight - 1) - 2);
Sort(x, 4);
*(ucDstImg + (iHeight - 0)*iWidth - 1) = x[2];
return 0;
}
三.均值化
int HistoNormalize(unsigned char* ucImg,unsigned char* ucNormImg, int iWidth,int iHeight) {
unsigned int Histogram[256];
memset(Histogram, 0, 256 * sizeof(int));
for (int i = 0; i < iHeight; i++) {
for (int j = 0; j < iWidth; j++) {
Histogram[ucImg[i*iWidth + j]]++;
}
}
double dMean = 0;
for (int i = 1; i < 255; i++) {
dMean += i * Histogram[i];
}
dMean = int(dMean/(iWidth*iHeight));
double dSigma = 0;
for (int i = 0; i < 255; i++) {
dSigma += Histogram[i] * (i - dMean)*(i-dMean);
}
dSigma /= (iWidth*iHeight);
dSigma = sqrt(dSigma);
double dMean0 = 128, dSigma0 = 128;
double dCoeff = dSigma0 / dSigma;
for (int i = 0; i < iHeight; i++) {
for(int j=0;j<iWidth;j++){
double dVal = ucImg[i*iWidth + j];
dVal = dMean0 + dCoeff * (dVal-dMean0);
if (dVal < 0) {
dVal = 0;
}
else if (dVal > 255) {
dVal = 255;
}
ucNormImg[i*iWidth + j] = (unsigned char)dVal;
}
}
return 0;
}
四.方向计算
int ImgDirection(unsigned char* ucImg,float* fDirc,int iWidth,int iHeight) {
const int SEMISIZ = 7;
int dx[SEMISIZ * 2 + 1][SEMISIZ * 2 + 1];
int dy[SEMISIZ * 2 + 1][SEMISIZ * 2 + 1];
float fx, fy;
memset(fDirc, 0, iWidth*iHeight * sizeof(float));
for (int y = SEMISIZ + 1; y < iHeight - SEMISIZ - 1; y++) {
for (int x = SEMISIZ + 1; x < iWidth - SEMISIZ - 1; x++) {
for (int j = 0; j < SEMISIZ * 2 + 1; j++) {
for (int i = 0; i < SEMISIZ * 2 + 1; i++) {
int index1 = (y + j - SEMISIZ)*iWidth + x + i - SEMISIZ;
int index2 = (y + j - SEMISIZ)*iWidth + x + i - SEMISIZ - 1;
int index3 = (y + j - SEMISIZ - 1)*iWidth + x + i - SEMISIZ;
dx[i][j] = int(ucImg[index1] - ucImg[index2]);
dy[i][j] = int(ucImg[index1] - ucImg[index3]);
}
}
fx = 0.0;
fy = 0.0;
for (int j = 0; j < SEMISIZ * 2 + 1; j++) {
for (int i = 0; i < SEMISIZ * 2 + 1; i++) {
fx += 2 * dx[i][j] * dy[i][j];
fy += (dx[i][j] * dx[i][j] - dy[i][j] * dy[i][j]);
}
}
fDirc[y*iWidth + x] = atan2(fx,fy);
}
}
return 0;
}
五.频率计算
int DircLowPass(float *fDirc,float* fFitDirc,int iWidth,int iHeight) {
const int DIR_FILTER_SIZE = 2;
int blocksize = 2 * DIR_FILTER_SIZE + 1;
int imgsize = iWidth * iHeight;
float *filter = new float[blocksize*blocksize];
float *phix = new float[imgsize];
float *phiy = new float[imgsize];
float *phi2x = new float[imgsize];
float *phi2y = new float[imgsize];
memset(fFitDirc,0,sizeof(float)*iWidth*iHeight);
float tempSum = 0.0;
for (int y = 0; y < blocksize; y++) {
for (int x = 0; x < blocksize; x++) {
filter[y*blocksize + x] = (float)(blocksize - (abs(DIR_FILTER_SIZE - x) + abs(DIR_FILTER_SIZE - y)));
tempSum += filter[y*blocksize + x];
}
}
for (int y = 0; y < blocksize; y++) {
for (int x = 0; x < blocksize; x++) {
filter[y*blocksize + x] /= tempSum;
}
}
for(int y=0;y<iHeight;y++){
for (int x = 0; x < iWidth; x++) {
phix[y*iWidth + x] = cos(fDirc[y*iWidth + x]);
phiy[y*iWidth + x] = sin(fDirc[y*iWidth + x]);
}
}
memset(phi2x, 0, sizeof(float)*imgsize);
memset(phi2y, 0, sizeof(float)*imgsize);
float nx, ny;
int val;
for (int y = 0; y < iHeight - blocksize; y++) {
for (int x = 0; x < iWidth - blocksize; x++) {
nx = 0.0;
ny = 0.0;
for (int j = 0; j < blocksize; j++) {
for (int i = 0; i < blocksize; i++) {
val = (x + i) + (j + y)*iWidth;
nx += filter[j*blocksize + i] * phix[val];
ny += filter[j*blocksize + i] * phiy[val];
}
}
val = x + y * iWidth;
phi2x[val] = nx;
phi2y[val] = ny;
}
}
for (int y = 0; y < iHeight - blocksize; y++) {
for (int x = 0; x < iWidth - blocksize; x++) {
val = x + y * iWidth;
fFitDirc[val] = atan2(phi2y[val],phi2x[val])*0.5;
}
}
delete[] phi2y;
delete[] phi2x;
delete[] phiy;
delete[] phix;
return 0;
}
六.掩码计算
int GetMask(unsigned char* ucImg,float *fDirection,float *fFrequency,unsigned char *ucMask,int iWidth,int iHeight) {
float freqMin = 1.0 / 25.0;
float freqMax = 1.0 / 3.0;
int x, y, k;
int pos, posout;
memset(ucMask,0,iWidth*iHeight);
for (y = 0; y < iHeight; y++) {
for (x = 0; x < iWidth; x++) {
pos = x + y * iWidth;
posout = x + y * iWidth;
ucMask[posout] = 0;
if (fFrequency[pos] >= freqMin && fFrequency[pos] <= freqMax) {
ucMask[posout] = 255;
}
}
}
for (k = 0; k < 4; k++) {
for (y = 1; y < iHeight - 1; y++) {
for (x = 1; x < iWidth - 1; x++) {
if (ucMask[x + y * iWidth] == 0xFF) {
ucMask[x - 1 + y * iWidth] |= 0x80;
ucMask[x + 1 + y * iWidth] |= 0x80;
ucMask[x + (y-1) * iWidth] |= 0x80;
ucMask[x + (y+1) * iWidth] |= 0x80;
}
}
}
for (y = 1; y < iHeight - 1; y++) {
for (x = 1; x < iWidth - 1; x++) {
if (ucMask[x + y * iWidth]) {
ucMask[x + y * iWidth] = 0xFF;
}
}
}
}
for (k = 0; k < 12; k++) {
for (y = 1; y < iHeight - 1; y++) {
for (x = 1; x < iWidth - 1; x++) {
if (ucMask[x + y * iWidth] == 0x0) {
ucMask[x - 1 + y * iWidth] &= 0x80;
ucMask[x + 1 + y * iWidth] &= 0x80;
ucMask[x + (y - 1) * iWidth] &= 0x80;
ucMask[x + (y + 1) * iWidth] &= 0x80;
}
}
}
for (y = 1; y < iHeight - 1; y++) {
for (x = 1; x < iWidth - 1; x++) {
if (ucMask[x + y * iWidth] != 0xFF) {
ucMask[x + y * iWidth] = 0x0;
}
}
}
}
return 0;
}
七.Gabor增强
int GaborEnhance(unsigned char* ucImg, float* fDirection, float* fFrequency, unsigned char* ucMask, unsigned char* ucImgEnhanced, int iWidth, int iHeight) {
const float PI = 3.141592654;
int i, j, u, v;
int wg2 = 5;
float sum, f, g;
float x2, y2;
float dx2 = 1.0 / (4.0*4.0);
float dy2 = 1.0 / (4.0*4.0);
memset(ucImgEnhanced,0,iWidth*iHeight);
for (j = wg2; j < iHeight - wg2; j++) {
for (i = wg2; i < iWidth - wg2; i++) {
if (ucMask[i + j * iWidth] == 0) {
continue;
}
g = fDirection[i+j*iWidth];
f = fFrequency[i+j*iWidth];
g += PI / 2;
sum = 0.0;
for (v = -wg2; v <= wg2; v++) {
for (u = -wg2; u <= wg2; u++) {
x2 = -u * sin(g) + v * cos(g);
y2 = u * cos(g) + v * sin(g);
sum += exp(-0.5*(x2*x2*dx2 + y2 * y2*dy2))*cos(2 * PI*x2*f)*ucImg[(i - u) + (j - v)*iWidth];
}
}
if (sum > 255.0) {
sum = 255.0;
}
if (sum < 0.0) {
sum = 0.0;
}
ucImgEnhanced[i + j * iWidth] = (unsigned char)sum;
}
}
return 0;
}
八.二值化
int BinaryImg(unsigned char* ucImage,unsigned char* ucBinImage,int iWidth,int iHeight,unsigned char uThreshold) {
unsigned char *pStart = ucImage, *pEnd = ucImage + iWidth * iHeight;
unsigned char *pDest = ucBinImage;
while (pStart < pEnd) {
*pDest = *pStart > uThreshold ? 1 : 0;
pStart++;
pDest++;
}
return 0;
}
int BinaryToGray(unsigned char *ucBinImg,unsigned char *ucGrayImg,int iWidth,int iHeight) {
unsigned char *pStart = ucBinImg, *pEnd = ucBinImg + iWidth * iHeight;
unsigned char *pDest = ucGrayImg;
while (pStart<pEnd) {
*pDest = (*pStart) > 0 ? 255 : 0;
pStart++;
pDest++;
}
return 0;
}
九.细化
int Thinning(unsigned char *ucBinedImg,unsigned char *ucThinnedImage,int iWidth,int iHeight,int iIterativeLimit) {
unsigned char x1, x2, x3, x4, x5, x6, x7, x8, xp;
unsigned char g1, g2, g3, g4;
unsigned char b1, b2, b3, b4;
unsigned char np1, np2, npm;
unsigned char *pUp, *pDown, *pImg;
int iDeletePoints = 0;
memcpy(ucThinnedImage,ucBinedImg,iWidth*iHeight);
for (int it = 0; it < iIterativeLimit; it++) {
iDeletePoints = 0;
for (int i = 1; i < iHeight - 1; i++) {
pUp = ucBinedImg + (i - 1)*iWidth;
pImg = ucBinedImg + i * iWidth;
pDown = ucBinedImg + (i + 1)*iWidth;
for (int j = 1; j < iWidth - 1; j++) {
pUp++;
pImg++;
pDown++;
if (!*pImg) {
continue;
}
x6 = *(pUp - 1);
x5 = *(pImg - 1);
x4 = *(pDown - 1);
x7 = *pUp;
xp = *pImg;
x3 = *pDown;
x8 = *(pUp + 1);
x1 = *(pImg+1);
x2 = *(pDown+1);
b1 = !x1 && (x2 == 1 || x3 == 1);
b2 = !x3 && (x4 == 1 || x5 == 1);
b3 = !x5 && (x6 == 1 || x7 == 1);
b4 = !x7 && (x8 == 1 || x1 == 1);
g1 = (b1 + b2 + b3 + b4) == 1;
np1 = x1 || x2;
np1 += x3 || x4;
np1 += x5 || x6;
np1 += x7 || x8;
np2 = x2 || x3;
np2 += x4 || x5;
np2 += x6 || x7;
np2 += x8 || x1;
npm = np1 > np2 ? np2 : np1;
g2 = npm >= 2 && npm <= 3;
g3 = (x1 && (x2 || x3 || !x8)) == 0;
g4 = (x5 && (x6 || x7 || !x4)) == 0;
if (g1&&g2&&g3) {
ucThinnedImage[iWidth*i + j] = 0;
++iDeletePoints;
}
}
}
memcpy(ucBinedImg,ucThinnedImage,iWidth*iHeight);
for (int i = 1; i < iHeight - 1;i++) {
pUp = ucBinedImg + (i - 1)*iWidth;
pImg = ucBinedImg + i * iWidth;
pDown = ucBinedImg + (i+1) * iWidth;
for (int j = 1; j < iWidth - 1; j++) {
pUp++;
pImg++;
pDown++;
if (!*pImg) {
continue;
}
x6 = *(pUp - 1);
x5 = *(pImg - 1);
x4 = *(pDown - 1);
x7 = *pUp;
xp = *pImg;
x3 = *pDown;
x8 = *(pUp + 1);
x1 = *(pImg + 1);
x2 = *(pDown + 1);
b1 = !x1 && (x2 == 1 || x3 == 1);
b2 = !x3 && (x4 == 1 || x5 == 1);
b3 = !x5 && (x6 == 1 || x7 == 1);
b4 = !x7 && (x8 == 1 || x1 == 1);
g1 = (b1 + b2 + b3 + b4) == 1;
np1 = x1 || x2;
np1 += x3 || x4;
np1 += x5 || x6;
np1 += x7 || x8;
np2 = x2 || x3;
np2 += x4 || x5;
np2 += x6 || x7;
np2 += x8 || x1;
npm = np1 > np2 ? np2 : np1;
g2 = npm >= 2 && npm <= 3;
g3 = (x1 && (x2 || x3 || !x8)) == 0;
g4 = (x5 && (x6 || x7 || !x4)) == 0;
if (g1&&g2&&g4) {
ucThinnedImage[iWidth*i+j] = 0;
++iDeletePoints;
}
}
}
memcpy(ucBinedImg,ucThinnedImage,iWidth*iHeight);
if (iDeletePoints == 0) {
break;
}
}
for (int i = 0; i < iHeight; i++) {
for (int j = 0; j < iWidth; j++) {
if (i < 16) {
ucThinnedImage[i*iWidth + j] = 0;
}
else if (i >= iHeight - 16) {
ucThinnedImage[i*iWidth + j] = 0;
}
else if (j < 16) {
ucThinnedImage[i*iWidth + j] = 0;
}
else if (j >= iWidth - 16) {
ucThinnedImage[i*iWidth + j] = 0;
}
}
}
return 0;
}
十.特征提取
int CutEdge(MINUTIAE* minutiaes,int count,unsigned char*ucImg,int iWidth,int iHeight) {
int minuCount = count;
int x, y, type;
bool del;
int *pFlag = new int[minuCount];
memset(pFlag,0,sizeof(int)*minuCount);
for (int i = 0; i < minuCount; i++) {
y = minutiaes[i].y - 1;
x = minutiaes[i].x - 1;
type = minutiaes[i].type ;
del = true;
if (x < iWidth / 2) {
if (abs(iWidth / 2 - x) > abs(iHeight / 2 - y)) {
while (--x >= 0) {
if (ucImg[x + y * iWidth] > 0) {
del = false;
break;
}
}
}
else {
if (y > iHeight / 2) {
while (++y < iHeight) {
if (ucImg[x + y * iWidth] > 0) {
del = false;
break;
}
}
}
else {
while (--y == 0) {
if (ucImg[x + y * iWidth] > 0) {
del = false;
break;
}
}
}
}
}
else {
if (abs(iWidth / 2 - x) > abs(iHeight / 2 - y)) {
while (++x < iWidth) {
if (ucImg[x + y * iWidth] > 0) {
del = false;
break;
}
}
}
else {
if (y > iHeight / 2) {
while (++y < iHeight) {
if (ucImg[x + y * iWidth] > 0) {
del = false;
break;
}
}
}
else {
while (--y >= 0) {
if (ucImg[x + y * iWidth] > 0) {
del = false;
break;
}
}
}
}
}
if (del) {
pFlag[i] = 1;
continue;
}
}
int newCount = 0;
for (int i = 0; i < minuCount; i++) {
if (pFlag[i] == 0) {
memcpy(&minutiaes[newCount],&minutiaes[i],sizeof(MINUTIAE));
newCount++;
}
}
delete[] pFlag;
pFlag = NULL;
return newCount;
}
十一.特征过滤
int MinuFilter(unsigned char *minuData,unsigned char *thinData,MINUTIAE *minutiaes,int &minuCount,int iWidth,int iHeight) {
float *dir = new float[iWidth*iHeight];
memset(dir,0,iWidth*iHeight*sizeof(float));
ImgDirection(thinData,dir,iWidth,iHeight);
unsigned char* pImg;
unsigned char val;
int temp = 0;
for (int i = 1; i < iHeight - 1; i++) {
pImg = minuData + i * iWidth;
for (int j = 1; j < iWidth - 1; j++) {
++pImg;
val = *pImg;
if(val>0){
minutiaes[temp].x = j + 1;
minutiaes[temp].y = i + 1;
minutiaes[temp].theta = dir[i*iWidth+j];
minutiaes[temp].type = int(val);
++temp;
}
}
}
delete[] dir;
minuCount = CutEdge(minutiaes,minuCount,thinData,iWidth,iHeight);
int *pFlag = new int[minuCount];
memset(pFlag,0,sizeof(int)*minuCount);
int x1, x2, y1, y2, type1, type2;
for (int i = 0; i < minuCount; i++) {
x1 = minutiaes[i].x;
y1 = minutiaes[i].y;
type1 = minutiaes[i].type;
for (int j = i + 1; j < minuCount; j++) {
if (pFlag[i] == 1) {
continue;
}
x2 = minutiaes[j].x;
y2 = minutiaes[j].y;
type2 = minutiaes[j].type;
int r = (int)sqrt(float((y1-y2)*(y1-y2)+(x1-x2)*(x1-x2)));
if (r <= 4) {
if (type1 == type2) {
if (type1 == 1) {
pFlag[i] = pFlag[j] = 1;
}
else {
pFlag[j] = 1;
}
}
else if (type1 == 1) {
pFlag[i] = 1;
}
else {
pFlag[j] = 1;
}
}
}
}
int newCount = 0;
for (int i = 0; i < minuCount; i++) {
if (pFlag[i] == 0) {
memcpy(&minutiaes[newCount],&minutiaes[i],sizeof(MINUTIAE));
newCount++;
}
}
delete[] pFlag;
minuCount = newCount;
return 0;
}
基于MFC开发的指纹识别系统.的更多相关文章
- 基于STM32单片机光学指纹识别模块（FPM10A）全教程(基于C语言)
本文转载,其来源在参考中:1,稍加修改,因为近期使用到这个模块,故而加以整理! 1.平台 首先我使用的是 奋斗 STM32 开发板 MINI板 基于STM32单片机光学指纹识别模块(FPM10A)全教 ...
- 基于Hadoop开发网络云盘系统客户端界面设计初稿
基于Hadoop开发网络云盘系统客户端界面设计初稿 前言: 本文是<基于Hadoop开发网络云盘系统架构设计方案>的第二篇,针对界面原型原本考虑有两个方案:1.类windows模式,文件夹 ...
- 基于Hadoop开发网络云盘系统架构设计方案
基于Hadoop开发网络云盘系统架构设计方案第一稿 引言 云计算技术的发展,各种网络云盘技术如雨后春笋,层出不穷,百度.新浪.网易都推出了自己的云盘系统,本文基于开源框架Hadoop设计实现了一套自己 ...
- 基于深度学习的人脸识别系统（Caffe+OpenCV+Dlib）【一】如何配置caffe属性表
前言 基于深度学习的人脸识别系统,一共用到了5个开源库:OpenCV(计算机视觉库).Caffe(深度学习库).Dlib(机器学习库).libfacedetection(人脸检测库).cudnn(gp ...
- 基于深度学习的人脸识别系统（Caffe+OpenCV+Dlib）【三】VGG网络进行特征提取
前言 基于深度学习的人脸识别系统,一共用到了5个开源库:OpenCV(计算机视觉库).Caffe(深度学习库).Dlib(机器学习库).libfacedetection(人脸检测库).cudnn(gp ...
- 基于深度学习的人脸识别系统（Caffe+OpenCV+Dlib）【二】人脸预处理
前言 基于深度学习的人脸识别系统,一共用到了5个开源库:OpenCV(计算机视觉库).Caffe(深度学习库).Dlib(机器学习库).libfacedetection(人脸检测库).cudnn(gp ...
- 基于深度学习的人脸识别系统系列（Caffe+OpenCV+Dlib）——【四】使用CUBLAS加速计算人脸向量的余弦距离
前言 基于深度学习的人脸识别系统,一共用到了5个开源库:OpenCV(计算机视觉库).Caffe(深度学习库).Dlib(机器学习库).libfacedetection(人脸检测库).cudnn(gp ...
- 基于HTML Canvas实现“指纹识别”技术
https://browserleaks.com/canvas 说明所谓指纹识别是指为每个设备标识唯一标识符(以下简称UUID).诸如移动原生的APP都可以通过调用相关设备API来获取相应的UUID. ...
- 基于TensorFlow的车牌号识别系统
简介 过去几周我一直在涉足深度学习领域,尤其是卷积神经网络模型.最近,谷歌围绕街景多位数字识别技术发布了一篇不错的paper.该文章描述了一个用于提取街景门牌号的单个端到端神经网络系统.然后,作者阐述 ...
随机推荐
- 2018-2019-2 20165336《网络对抗技术》Exp0 Kali安装 Week1
2018-2019-2 20165336<网络对抗技术>Exp0 Kali安装 Week1 一.选择官网kali linux系统的版本 二.配置虚拟机 根据 安装教程(https://bl ...
- nginx修改上传文件大小限制
问题: 项目上线,图片上传报413错误,找了半天,原来是nginx限制了上传大小 在nginx.conf的server的location中加client_max_body_size 10m;
- PyQt5之使用Qt下的designer工具将.ui文件转换成.py文件后添加什么东西后方可运行
首先证明我是加了那些鬼东西以后可以成功运行的. 然后来叙述一下我的过程. 这是一个.ui文件生成的.py文件.(把主要的内容省去了,但是没有影响结构) # -*- coding: utf-8 -*- ...
- matlab工作空间数据导入simulink
使用的是其中一种方式: 第一步在工作命令区 ,写命令: 第二步:保证导入simulink区,及from worker设置: 其中注意设置你的采样时间, 第三步设置scop : 采样时承接数据线上 ...
- 【Nginx】-NO.141.Nginx.1 -【Nginx】
Style:Mac Series:Java Since:2018-09-10 End:2018-09-10 Total Hours:1 Degree Of Diffculty:5 Degree Of ...
- FG面经Prepare: BST to Double LinkedList
BST to double linkedlist in inorder traversal sequenceFollow Up: 如果这个BST自带prev, next, 给一个value,插进去这个 ...
- Java多线程处理List数据
实例1: 解决问题:如何让n个线程顺序遍历含有n个元素的List集合 import java.util.ArrayList; import java.util.List; import org.apa ...
- EF：分页查询 + 条件查询 + 排序
/// <summary> /// linq扩展类---zxh /// </summary> /// <typeparam name="T">& ...
- （转）批量插入sql语句
为了减少数据库连接的I/O开销,一般会把多条数据插入放在一条SQL语句中一次执行.1.INSERT INTO TABLE(col1, col2) VALUES(val11, val12), (val2 ...
- ls file less
ls-a 列出所有文件,包含隐藏文件-l 以长格式显示结果-F 在所列出的文件后面显示其格式-r 按照文件名降序展示-t 按照时间列出-S 按照文件大小排序 file 文件名:展示文件的类型简单描述 ...