haoyGONG/LPDGAN

Train test parition

Opened this issue · 1 comments

In your paper you state there are 9288 + 1000 = 10.288 image pairs. However, the dataset provided has 10.408 pairs. Why are there more? And which is the exact partition?

The extra images were collected after the completion of the paper. The data used in the experiment in the original paper were divided into different sets based on the first 10288 images in ascending number order. The additional 120 images were collected subsequently.