intelpro / uevd_public Goto Github PK

Hi,

congratulations for your amazing research work!

Do you plan to open-source the GoPro datasets + simulated motion blur + simulated events? This would make comparison to your Table 2 possible.

If not, can you please specifify a bit more details about how to generate the GoPro dataset? E.g. which constrast threshold is used? You mention "We synthesize blurry video by averaging the video frames. To mimic the real video frame
acquisition, we follow the method in [55]. " How do you average the frames? Do you use the code from [55] for that?

Thanks

Best
Simon

This work is an interesting work. And could you privide the pre-trained models? I am interested in running this method with the pre-trained models on my own dataset.