GithubHelp home page GithubHelp logo

Comments (8)

nkolot avatar nkolot commented on July 22, 2024

I'm pretty sure that it has to do with the fact that the renderer currently does not support loading obj files containing vertex normals.

@hiroharu-kato do you think there is an easy way of fixing it?

from neural_renderer.

qilimk avatar qilimk commented on July 22, 2024

Thanks for replying so quickly, I tried to delete all vertex normals lines and its preview is still normal, but result from renderer is also no-changed. Do you think it is because there are too many groups in the obj file?

from neural_renderer.

nkolot avatar nkolot commented on July 22, 2024

Have you tried rendering the silhouette only? How do the results look like?

from neural_renderer.

nkolot avatar nkolot commented on July 22, 2024

I tried to render the silhouettes and they look correct. I also tried to renderer it using a uniform texture (ones everywhere) and it looks as it should do. I am seeing the same results as you. Can you try to run your examples with the original neural renderer? Do you get the same results?

from neural_renderer.

qilimk avatar qilimk commented on July 22, 2024

Sorry for ambiguous expression. I didn't render the silhouettes, and I think you are right that the renderer does not support the vertex normals, so I delete all the vertex normals lines and make a new obj file. The result is below.

camaro_no_vn_comp
Here is the new obj and mtl files.
The Preview(the Preview program on Mac, which uses OpenGL renderer) result is still different from rendering result.

from neural_renderer.

nkolot avatar nkolot commented on July 22, 2024

I tried to render it using different software, MacOS Preview, Meshlab and Cinema4D and I am seeing completely different visualization results in each one. This is really annoying, so I can't come up with an easy fix for that.

from neural_renderer.

qilimk avatar qilimk commented on July 22, 2024

Thanks so much, I may try some other solutions! If you have any new idea to solve this problem, could you let me know about it?

from neural_renderer.

czw0078 avatar czw0078 commented on July 22, 2024

It seems that those 3D software editors do not even have a consistency with each other on obj. file format. To solve this issue systematically, see this mini cooper again. Note the second line of raw file

# Exported from SketchUp, (c) 2000-2012 Trimble Navigation Limited

This 3d file is actually converted from .skp file to .obj by SketchUp. The 3D Warehouse (supported by the same company developing SketchUp) has a very large repo for all kinds of free 3d objects. Many researchers build 3d object databases by collecting models from 3D Warehouse.

So, I think perhaps we should try to make it compatible with .obj files converted from .skp files. ( Or even develop a new python code load_skp.py just like load_obj.py, directly reads out from .skp file. The .skp file format seems a little bit more human readable, but it is still a lot of work if decide to do it I guess.)

from neural_renderer.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.