Occluded Point Map

Description

This repository contains two general functions where it is possible to get:

Function	Input	Result
A projection of mesh to equirectangular image (JPG file): projection_mesh_to_image	360 image (1024x512 px) Boundary Mesh	Projected Mesh to 360º image
A occluded point map from mesh and furniture (PLY file): occluded_point_map	Boundary Mesh + Furniture	Occluded Point Map Number of occluded points Statistics (CSV file) statistics.csv

Quick start
Source Code Structure
Installation

Quick start

Run next command line:

$ make quick-start

Source Code Structure

├── colorized_mesh
|   ├── colorized_mesh.py
|   ├── obb_tree.py
├── dataloader
|   ├── mesh_dataset.py
├── dataset
|   ├── [...]
|   ├── README.md
├── docs
├── mesh
│   ├── manager.py
│   ├── projection_img_mesh.py
├── misc
│   ├── panostretch.py
│   ├── point2mesh.py
|   ├── tools.py
├── result
|   ├── README.md
├── settings
│   ├── argument_parser.py
│   ├── arguments.py
|   ├── config_files
|       ├── local_config.ini
├── utils
│   ├── save_results.py
├── .gitignore
├── main.py
├── Makefile
├── README.md
└── requirements.txt

Installation

Virtual Environment (Optional)
Requirements

Virtual Environment (Optional)

Install pip3 if it is already not installed on your system:

$ sudo apt install python3-pip

Install Python3-venv if it is already not installed on your system:

$ sudo apt install -y python3-venv

To create a virtual environment, decide upon a directory where you want to place it, and run the venv module as a script with the directory path (Note: the second venv is the name of your virtual environment, so you can choose the one you want):

$ python3 -m venv venv

Once you’ve created a virtual environment, you may activate it and run it:

$ source venv/bin/activate

Install the requirements file:

(venv) $ pip3 install -r requirements.txt

Requirements

This code was tested with the following tools:

Python 3.8.10
Open3D 0.13.0 (which includes the installation of Numpy and more)
Trimesh 3.9.27
Pytorch3D 0.3.0 (which includes the installation of torchvision)
Vedo 2021.0.3
VTK 9.0.3
PyVista 0.31.3

Install all the requirements for this project:

pip3 install -r requirements.txt

Installing torch==1.9.0

pip3 install torch

Check the requirements:

pip3 freeze

Settings

The configuration files are in the /config_files folder and the name is local_config.ini.

cat /settings/config_files/local_config.ini

The default values are indicated by 'DEFAULT'. Example:

[DEFAULT]
img_path = /Datasets/img
mesh_path = /Datasets/mesh
index = 0

Concepts

Triangle mesh

Reference

A triangle mesh has several properties that can be tested:

the manifold property, where the triangle mesh has to be edge manifold and vertex manifold.
- A triangle mesh is edge manifold, if each edge is bounding either one or two triangles.
- Triangle mesh is vertex manifold if the star of the vertex is edge-manifold and edge-connected, e.g., two or more faces connected only by a vertex and not by an edge.
Another property is the self-intersection.
A watertight mesh can be defined as a mesh that is edge manifold, vertex manifold and not self intersecting.
if it is orientable, i.e. the triangles can be oriented in such a way that all normals point towards the outside.

More info

There is a function to check those properties in this project here.

EvaAlmansa/ColorizedMesh