Train and test split of dataset

[LuoBingjun]

Thank you for your excellent work.
I'm trying to reproduce the results in your paper. I have successfully generated 221 output images on RoadScene. But I find the evaluation metrics on the whole RoadScene dataset are not completely consistent to the paper:

| EN | SD | SF | MI | SCD | VIF | Qabf | SSIM |
| 7.19 | 40.95 | 12.13 | 1.97 | 1.71 | 0.57 | 0.47 | 0.98 |

Can you provide the train and test split settings of datasets in your experiments?

[Wbzb]

I have the same question about the evaluation on the TNO dataset. SD shown in the paper is 51.42 but when I use the provided DDFM code I got SD is 34.55

[zhongyuan1996]

Thanks for providing the code. I have also tried the RoadScene dataset, with crop_LR_vis as the visible image, and ir as the infrared image. The average Entropy I obtained is 7.1922, and the standard deviation is 41.4473. Although I did not calculate other metrics, I would assume my output is more consistent with @LuoBingjun 's result.