rendering with pbrt about tof_rgbd_processing

[luckyluliu]

What do you mean by pbrt-tof, seems you only provide a 'more.pbrt', no 'more-tof.pbrt' found. What's the difference between more.pbrt and more-tof.pbrt??? @sylqiu

[sylqiu]

Hi, in the command line you seem to be using the original pbrt binary instead of the pbrt-tof (or something named like this) binary. Can you double check which binary file is used. Let me know if this solves the problem.

…

On 21 Aug 2020, at 23:00, Junggy ***@***.***> wrote:  I have followed your instruction, but I was not able to get 256 transient images (i rather got single *.exr file) I downloaded ToF-pbrt-generation and installed pbrt-v3-tof when I run pbrt, I used following command /DirToPbrt/pbrt /DirToToF/ToF-pbrt-generation/scenes/ToFFlyingThings3D/breakfast/more.pbrt and I got single "breakfast.exr" file in /DirToPbrt/tri Also I tried to use "batch_run_pbrt.sh" but I was not able to run it as I have no clue about some variables to set like CAMFILE (I used ToF-pbrt-generation/scenes/ToFFlyingThings3D/breakfast/campath1000.txt (or campath1000_z.txt)) , MAT_PATH (I have no idea), SCENE_LIST (also have no idea) etc can I get more detailed instruction of way to get 256 transient images? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[Junggy]

@sylqiu
So, I was able to render 256 transient images from one viewpoint

But when I get depth from those images by using "transient_to_depth.m"
i.e. [phase, depth] = transient_to_depth('./', 240, 320, 256, (40:10:70)*1e6, 10);

resulting depth image doesn't seem to be promising.
It was rather very noisy and looks rather like amplitude image

(i.e. camera is looking at the corner of the room and some flying objects are in front of camera

• wall on the right, intensity increases as wall gets further
• wall on the left is way darker but intensity is nearly static
• object right in front of the camera is partially very bright

which doesn't seems to be correct depth.)

and correlation images are also a bit strange

(i.e. - first correlation image is rather white background and little black noise

• second correlation image looks very close to the depth image i got
• third correlation image is rather white background with intense black noise
• fourth correlation image seems to be quite over-saturated version of the first correlation)

maybe Issue I have is that either from 256 transient image generation or (mostly) from correlation calculation.

is there any good way to debug this?
or is it possible to share like one set of 256 transient images that I can check?

[sylqiu]

The phase images don't look correct. I think you may need to check two things. - play the transient volume as a video to see if there are "light in flight", to make sure you get the correct transient rendering result - find (or implement yourself) the code that retrieve of a single phase without phase unwrapping. that phase image should at least look like a depth image except it will look periodic in the z-dimension. Then check the code in transient_to_depth for the phase unwrapping part.

…

On Thu, Aug 27, 2020 at 11:31 PM Junggy ***@***.***> wrote: So, I was able to render 256 transient images from one viewpoint But when I get depth from those images by using "transient_to_depth.m" i.e. [phase, depth] = transient_to_depth('./', 240, 320, 256, (40:10:70)*1e6, 10); resulting depth image doesn't seem to be promising. It was rather very noisy and looks rather like amplitude image (i.e. camera is looking at the corner of the room and some flying objects are in front of camera - wall on the right, intensity increases as wall gets further - wall on the left is way darker but intensity is nearly static - object right in front of the camera is partially very bright [image: correlation] <https://user-images.githubusercontent.com/45639167/91462871-fb737480-e88a-11ea-9ad2-bb2066bcc9cd.png> [image: depth] <https://user-images.githubusercontent.com/45639167/91462876-fd3d3800-e88a-11ea-82b2-4d4e76e29a20.png> which doesn't seems to be correct depth.) and correlation images are also a bit strange (i.e. - first correlation image is rather white background and little black noise - second correlation image looks very close to the depth image i got - third correlation image is rather white background with intense black noise - fourth correlation image looks very close to the depth image I got) maybe Issue I have is that either from 256 transient image generation or (mostly) from correlation calculation. is there any way that I can debug about what is wrong? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#7 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AF3DRX2Y6R74GDXCAMU446LSCZ34LANCNFSM4QHL4VMQ> .

-- Best regards, Syl

[luckyluliu]

I get the same result of depth with yours on breakfast dataset, have you find out what is wrong??? @sylqiu @Junggy