The metrics provided in NTED paper is inconsistent with the ones in GFLA

[mlyarthur]

I note that the quantitative results in your new NTED paper is inconsistent with GFLA,can you explain this and point out the difference in your metrics calculation.
NTED

GFLA

[RenYurui]

Hi,
Thanks for your interest. This difference is caused by the evaluation images. In this paper, we use the resized High-Resolution images in DeepFashion as our evaluation images.
The FID is calculated by comparing the distribution of the generated images with the distribution of the images in the training set.
The SSIM and LPIPS are calculated between the generated images and the corresponding GT images.
We will provide the evaluation scripts as soon as possible.

[ankanbhunia]

Thanks for the excellent work @RenYurui. It'd be really nice if you can share the evaluation scripts. Also, if I understood correctly the Table.1 is for the pose control. Did you also calculate the scores for appearance control?

[mlyarthur]

@RenYurui hi， how you resize the original high-resolution images? bilinear or bicubic