alignment/crop/resize/equalize etc

asked 2016-06-26 07:32:06 -0500

atv gravatar image

So things like cropping, alignment, grayscale and histogram modification as described in :

Does that apply to both the image i train on and the frame i get from the webcam ?

Currently i only convert it to grayscale, and resize it, and i get pretty good results (although the prediction fluctuates from 40-300 but that might depend on the quality of images i have)

What code would i need for this? Currently i only have this, and it's done on the webcam frame

cvtColor(original, gray, CV_BGR2GRAY); cv::resize(face, face_resized, Size(im_width, im_height), 1.0, 1.0, INTER_CUBIC);

I guess i should add this: equalizeHist(input,Equalizedimg); what else? Are those python scripts translatable to c++ code ?

edit retag flag offensive close merge delete