About training setting

[Rayn-Wu]

Hi, thanks for sharing this project.

In your paper, it is said that:

 "Sequence ”2014-12-09-13-21-02” and ”201412-16-18-44-24” are used for day-time and night-time training, respectively."

However, in Sec. 4.2:

First, we use CycleGAN[44] to translate day-time images to night-time images, then generated images and the day-time images are regarded as image pairs which are the input of the proposed domain-separated network.

Did you use the night-time data in training? Or just day-time + fake night-time(GAN)