What is the main idea of separating ppl on image to main person and other persons?
I notices in all places inside the code handling of this is absolutely same.
Every piece of code with use of joint_self followed by exactly same code with joint_others.
The 2 differences i've found:
1) main person placed close to center(random shift added later) and resized approximately to same size(random resizing is off)
2) This part of code which I don't really understand
// label size is image size/ stride
if (mode > 4){
for (int g_y = 0; g_y < grid_y; g_y++){
for (int g_x = 0; g_x < grid_x; g_x++){
for (int i = 0; i < np; i++){
float weight = float(mask_miss_aug.at<uchar>(g_y, g_x)) /255; //mask_miss_aug.at<uchar>(i, j);
// NEXT LINE IS VERY STRANGE //
if (meta.joint_self.isVisible[i] != 3){
transformed_label[i*channelOffset + g_y*grid_x + g_x] = weight;
}
}
// background channel
if(mode == 5){
transformed_label[np*channelOffset + g_y*grid_x + g_x] = float(mask_miss_aug.at<uchar>(g_y, g_x)) /255;
}
if(mode > 5){
transformed_label[np*channelOffset + g_y*grid_x + g_x] = 1;
transformed_label[(2*np+1)*channelOffset + g_y*grid_x + g_x] = float(mask_all_aug.at<uchar>(g_y, g_x)) /255;
}
}
}
}
Looks like if main person has one of the limb invisible, whole mask for that limb is not initialised i.e. limbs of other ppl are ingored.
What is the main idea of separating ppl on image to main person and other persons?
I notices in all places inside the code handling of this is absolutely same. Every piece of code with use of joint_self followed by exactly same code with joint_others.
The 2 differences i've found: 1) main person placed close to center(random shift added later) and resized approximately to same size(random resizing is off) 2) This part of code which I don't really understand
Looks like if main person has one of the limb invisible, whole mask for that limb is not initialised i.e. limbs of other ppl are ingored.