What are the proprties for get the coordinate to crop the images ? about pytorch_retinaface HOT 2 CLOSED

@NarasimmanSaravana1994

rectangle is added as following:

cv2_imshow(cv2.rectangle(im,(x,y),(x+w,y+h),(0,255,0),2))

I'd suggest numpy-approach for cropping, as:

crop_img = raw_img[y:y+h, x:x+w]
In dets object, we have b[0],b[1],b[2],b[3], where b[1], b[3] are y, y+h, and b[0], b[2] are x,x+w

Therefore, cropping goes as:

cropped = img_raw[b[1]:b[3], b[0]:b[2]]

Works like a charm!

from pytorch_retinaface.

You can refer to "_crop()" in https://github.com/biubug6/Pytorch_Retinaface/blob/master/data/data_augment.py for detials.
ps: It would calculate iou between box and crop_img.

from pytorch_retinaface.