This repository was created for evaluation at Industrial Vision Advanced Systems "SAVI 23-24 Trabalho prático 2".
Make Suggestion
Table of Contents
This assignment was developed for Advanced Systems of Industrial Vision. It was a project to learn point cloud and image manipulation as well as classifier Neural Networks.
This project is capable of extracting objects from a Point Cloud of a scene, classify them and extract their features, such as height or color. The model was trained with the Washington Dataset, so it is better suited for classifying household objects.
Before anything, the user should set up a environment variable:
$SAVI_TP2 --> /path_to_project
To run the program, the following libraries should be installed:
opencv
torch
torchmetrics
vlc_python
gTTs
open3D
matplotlib
numpy
To run the project, execute the following script under ./src
./main.py
Arguments when running main.py
:
- -s -> Choose which scene to render;
- -pNN -> Enable Neural Network debugging mode.
- -us -> Use upscale for the NN inputs (takes a lot more time)
Bruno Silva - bruno.favs@ua.pt
Mário Vasconcelos - mario.vasconcelos@ua.pt
- Professor Miguel Oliveira - mriem@ua.pt