Replies: 7 comments 7 replies
-
|
Hi! I am wondering about the method to learn global depth ordering. Did you try to learn a translation vector for each person and a single camera for the image? (i.e. 3N + 3 parameters, in contrast to current 3N parameters from camera map where N is the number of people) Actually SMPL encodes a prior and I think this method should work? @Arthur151 |
Beta Was this translation helpful? Give feedback.
-
|
Global depth ordering is intereting but challenge. It could be easier to estimate the depth order index (as you mensioned before). If we try to retrieve the actual camera metrix, it could be a little bit harder to get accurate parameters. But it should be work with proper settings. |
Beta Was this translation helpful? Give feedback.
-
Alright, I am gonna try this with your training code. Looking forward to 🥇 :) |
Beta Was this translation helpful? Give feedback.
-
|
Taining code has been released and tested by developers. |
Beta Was this translation helpful? Give feedback.
-
|
Hey Arthur! I saw this posted in the CVPR2022 booted up a demo in under 30min today, very impressive! Think any performance improvements possible if we're able to provide known camera parameters? Also, how to optimize to run on GPU (instead of just CPU)? OR even to use all my cores on CPU. Great work! First image segmentation I've seen that comes close to what my brain is able to infer from a single monocular. |
Beta Was this translation helpful? Give feedback.
-
|
Hi, Alex, Maybe deploy the model with ONNX / TensorRT would be helpful. |
Beta Was this translation helpful? Give feedback.
-
|
Hey Arthur! Great job for the pose extraction, excellent task in visualization of the 3d locations, a question to ask, by giving the function: romp --mode=webcam --show , is that possible to change the character and the background, the romp -h for help, does not show relevant information? |
Beta Was this translation helpful? Give feedback.
-
|
I think this is what you need. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you very much, Arthur, thumbs up, I tried the above repo, and it works, it is a pity that it does not apply the --show_largest function, for example, if there are multiple in front of the camera, the Armature is keeping on jumping from one to another. |
Beta Was this translation helpful? Give feedback.
-
|
Hi,Arthur.During my inference for trace.Something wrong occoured. PermissionError: [Errno 13] Permission denied: '/home/yusun' |
Beta Was this translation helpful? Give feedback.
-
|
There are some hardcoded paths inside of this project. Search for '/home/yusun' in all the python code. Replace it with your home directory. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you!
Alright done.Really good.
发自我的iPhone
…------------------ Original ------------------
From: Greg Cockroft ***@***.***>
Date: Tue,Jun 25,2024 7:43 PM
To: Arthur151/ROMP ***@***.***>
Cc: WUchuhan077 ***@***.***>, Comment ***@***.***>
Subject: Re: [Arthur151/ROMP] Welcome to ROMP Discussions! (Discussion #16)
There are some hardcoded paths inside of this project. Search for '/home/yusun' in all the python code. Replace it with your home directory.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you commented.Message ID: ***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
|
Hi,Arthur.When I downloaded the h36m files from Baidu network, I noticed that these are some files missing. How could I download the missing files? |
Beta Was this translation helpful? Give feedback.
-
|
There are only images.z02 03 04 05 06 07. The others are missing. |
Beta Was this translation helpful? Give feedback.
-
👋 Welcome!
We’re using Discussions as a place to connect with other members of our community. We hope that you:
build together 💪.
Beta Was this translation helpful? Give feedback.
All reactions