The format of outputs

[yuchenweng-private]

Hi, Shiba, Thank you for your excellent work. I have tried to implement my own dataloader to try your code on my dataset, it seems already worked, but I have outputs like '0.png', '1.png', '2.png', '3.png' that I don't know which one is iwe, which one is optical flow. From the generated images that I guess 3 is the optical flow, but which one is iwe generated by estimated optical flow?

[shiba24]

Hi Yuchen, nice to hear that you implement your own dataloader.

The 0.png etc is coming from my visualizer's naming convention.

• visualize_one_batch_warp is called twice, without motion and with motion here:

https://github.com/tub-rip/event_based_optical_flow/blob/main/main.py#L178-L180

• function: https://github.com/tub-rip/event_based_optical_flow/blob/main/src/solver/patch_contrast_pyramid.py#L518-L536

My visualizer incrementally name the save file. You can optionally specify the prefix using file_prefix args https://github.com/tub-rip/event_based_optical_flow/blob/main/src/visualizer.py#L158

Hope this helps. Let me know if the result do not look good. I could help digging into the flow result.

[shiba24]

@yuchenweng-private please let me know if you need more help. I close this issue for now as the original quesion is answered.