search

chungyiweng / humannerf

HumanNeRF turns a monocular video of moving people into a 360 free-viewpoint video.
MIT License
786 stars 86 forks source link

How to manually revise the human segmentation? #62

Open seunguk-do opened 1 year ago

seunguk-do commented 1 year ago

Hi, thank you for sharing great work. I came from the issue #10. Could you tell me how you revised the human segmentation? Are there any useful tools for it? It will be really helpful if you share some tips.

Dipankar1997161 commented 1 year ago

If you take a closer look at the zju-mocap dataset, the masks of every image is just pure black image of the same dimension as the input image. You can do this manually as well using OpenCV.

If this isn't your requirement, then try to generate silhouette images, there are many ways to do that. you can pick any.

I know I gave a very coarse description but that's how I did too. Let me know if anything is needed further.

mch0dmin commented 1 year ago

hi @Dipankar1997161 , I encountered the same confusion. Can you provide a detailed description of how you did it? Can you provide some reference codes? Thanks.

Dipankar1997161 commented 1 year ago

hi @Dipankar1997161 , I encountered the same confusion. Can you provide a detailed description of how you did it? Can you provide some reference codes? Thanks.

Hey @mch0dmin, Actually my project is slightly different, but what I can suggest you is download the dataset 313 and 315, in that there is a prepare_mask.py file, use that to generate the mask/Segmentation images needed.

It's only there in coreview_313 and 315 ( for Zju-mocap dataset) Hope it helps.

mch0dmin commented 1 year ago

hi @Dipankar1997161 , Thank you for your reply. I have resolved this issue. Thank you again.

Dipankar1997161 commented 1 year ago

@mch0dmin Glad to hear. Could you share it here, so it's useful for others too of its not confidential

2928254221 commented 1 year ago

你好@Dipankar1997161，我也遇到了同样的困惑。您能详细描述一下您是如何做到的吗？可以提供一些参考代码吗？谢谢。

嘿@mch0dmin， 实际上我的项目略有不同，但我可以建议您下载数据集313和315，其中有一个prepare_mask.py文件，用它来生成所需的掩模/分割图像。

它仅存在于 coreview_313 和 315（对于 Zju-mocap 数据集） 希望有帮助。

Where is the dataset 313 and 315？Any link?

Dipankar1997161 commented 1 year ago

你好@Dipankar1997161，我也遇到了同样的困惑。您能详细描述一下您是如何做到的吗？可以提供一些参考代码吗？谢谢。

嘿@mch0dmin， 实际上我的项目略有不同，但我可以建议您下载数据集313和315，其中有一个prepare_mask.py文件，用它来生成所需的掩模/分割图像。 它仅存在于 coreview_313 和 315（对于 Zju-mocap 数据集） 希望有帮助。

Where is the dataset 313 and 315？Any link?

It's in "Zju-mocap- dataset. https://github.com/zju3dv/neuralbody/blob/master/INSTALL.md#zju-mocap-dataset