search

microsoft / Deep3DFaceReconstruction

Accurate 3D Face Reconstruction with Weakly-Supervised Learning: From Single Image to Image Set (CVPRW 2019)
MIT License
2.18k stars 443 forks source link

Precompiled binary file rendered outputting all zeros #158

Closed gtangg12 closed 3 years ago

gtangg12 commented 3 years ago

Hi,

I am currently working on a docker workspace with Linux and TensorFlow 1.12.

Docker Linux version: DISTRIB_ID=Ubuntu DISTRIB_RELEASE=16.04 DISTRIB_CODENAME=xenial DISTRIB_DESCRIPTION="Ubuntu 16.04.5 LTS"

Docker Image: tensorflow/tensorflow:1.12.0-py3

Issue: I downloaded the precompiled file into my working dir and mounted my working dir onto the docker. I am able to run and generate a mesh. However, it appears that the rasterizer is outputting all zeros. In the code provided (rasterize_triangles.py), tmpp is directly sent to demo.py and printed, and the output is shown. I am fairly certain the pipeline I made is correct as when I change tmpp to vertices, it outputs the same values as face_mask. I also tried with vertex_ids and etc and it outputs 0s. Any fixes?

Thanks

Screen Shot 2021-05-16 at 6 09 41 AM Screen Shot 2021-05-16 at 6 10 27 AM
gtangg12 commented 3 years ago

My bad the edges are 0 due to no triangles.