void haralick2D(ImageMatrix &Im, double distance, double *out) {
unsigned int a,x,y;
unsigned char **p_gray;
TEXTURE *features;
int angle;
double min[14],max[14],sum[14];
double min_value,max_value;
double scale255;
readOnlyPixels pix_plane = Im.ReadablePixels();
if (distance <= 0) distance = 1;
p_gray = new unsigned char *[Im.height];
for (y = 0; y < Im.height; y++)
p_gray[y] = new unsigned char[Im.width];
Im.BasicStatistics(NULL, NULL, NULL, &min_value, &max_value, NULL, 0);
scale255 = (255.0/(max_value-min_value));
for (y = 0; y < Im.height; y++)
for (x = 0; x < Im.width; x++)
p_gray[y][x] = (unsigned char)((pix_plane(y,x) - min_value) * scale255);
for (a = 0; a < 14; a++) {
min[a] = INF;
max[a] = -INF;
sum[a] = 0;
}
for (angle = 0; angle <= 135; angle = angle+45) {
features = Extract_Texture_Features((int)distance, angle, p_gray, Im.height,Im.width);
/* (1) Angular Second Moment */
sum[0] += features->ASM;
if (features->ASM < min[0]) min[0] = features->ASM;
if (features->ASM > max[0]) max[0] = features->ASM;
/* (2) Contrast */
sum[1] += features->contrast;
if (features->contrast < min[1]) min[1] = features->contrast;
if (features->contrast > max[1]) max[1] = features->contrast;
/* (3) Correlation */
sum[2] += features->correlation;
if (features->correlation < min[2]) min[2] = features->correlation;
if (features->correlation > max[2]) max[2] = features->correlation;
/* (4) Variance */
sum[3] += features->variance;
if (features->variance < min[3]) min[3] = features->variance;
if (features->variance > max[3]) max[3] = features->variance;
/* (5) Inverse Diffenence Moment */
sum[4] += features->IDM;
if (features->IDM < min[4]) min[4] = features->IDM;
if (features->IDM > max[4]) max[4] = features->IDM;
/* (6) Sum Average */
sum[5] += features->sum_avg;
if (features->sum_avg < min[5]) min[5] = features->sum_avg;
if (features->sum_avg > max[5]) max[5] = features->sum_avg;
/* (7) Sum Variance */
sum[6] += features->sum_var;
if (features->sum_var < min[6]) min[6] = features->sum_var;
if (features->sum_var > max[6]) max[6] = features->sum_var;
/* (8) Sum Entropy */
sum[7] += features->sum_entropy;
if (features->sum_entropy < min[7]) min[7] = features->sum_entropy;
if (features->sum_entropy > max[7]) max[7] = features->sum_entropy;
/* (9) Entropy */
sum[8] += features->entropy;
if (features->entropy < min[8]) min[8] = features->entropy;
if (features->entropy > max[8]) max[8] = features->entropy;
/* (10) Difference Variance */
sum[9] += features->diff_var;
if (features->diff_var < min[9]) min[9] = features->diff_var;
if (features->diff_var > max[9]) max[9] = features->diff_var;
/* (11) Diffenence Entropy */
sum[10] += features->diff_entropy;
if (features->diff_entropy < min[10]) min[10] = features->diff_entropy;
if (features->diff_entropy > max[10]) max[10] = features->diff_entropy;
/* (12) Measure of Correlation 1 */
sum[11] += features->meas_corr1;
if (features->meas_corr1 < min[11]) min[11] = features->meas_corr1;
if (features->meas_corr1 > max[11]) max[11] = features->meas_corr1;
/* (13) Measure of Correlation 2 */
sum[12] += features->meas_corr2;
if (features->meas_corr2 < min[12]) min[12] = features->meas_corr2;
if (features->meas_corr2 > max[12]) max[12] = features->meas_corr2;
/* (14) Maximal Correlation Coefficient */
sum[13] += features->max_corr_coef;
if (features->max_corr_coef < min[13]) min[13] = features->max_corr_coef;
if (features->max_corr_coef > max[13]) max[13] = features->max_corr_coef;
free(features);
}
for (y = 0; y < Im.height; y++)
delete [] p_gray[y];
delete [] p_gray;
/* copy the values to the output vector in the right output order */
double temp[28];
for (a = 0; a < 14; a++) {
temp[a] = sum[a]/4;
temp[a+14] = max[a]-min[a];
}
out[ 0] = temp[ 0];
out[ 1] = temp[14];
out[ 2] = temp[ 1];
out[ 3] = temp[15];
out[ 4] = temp[ 2];
out[ 5] = temp[16];
out[ 6] = temp[ 9];
out[ 7] = temp[23];
out[ 8] = temp[10];
out[ 9] = temp[24];
out[10] = temp[ 8];
out[11] = temp[22];
out[12] = temp[11];
out[13] = temp[25];
out[14] = temp[ 4];
out[15] = temp[18];
out[16] = temp[13];
out[17] = temp[27];
out[18] = temp[12];
out[19] = temp[26];
out[20] = temp[ 5];
out[21] = temp[19];
out[22] = temp[ 7];
out[23] = temp[21];
out[24] = temp[ 6];
out[25] = temp[20];
out[26] = temp[ 3];
out[27] = temp[17];
}
