Results 1 comments of Bolun Wang

What we did in our experiments was to upsample all images to 224*224 (cause we need to fit it into VGG-Face). This should be fairly simple to do when you...