Open Redmond-AI opened 4 weeks ago
Hi, you can save the output depth video as a list of “.png” depth images, as converting images to video causes visual compression (please refer to here).
I'm not sure if you understand the artifacting I am talking about
I suppose a heavy median would fix this, but it would remove a lot of detail too. I modified the code to save 16bit pngs of "disparity". You can also see this artifacting in every example online.
Am I doing something wrong or is this expected behavior?
Is there any way to get rid of the "dithering" and export a smoother depth map?