YOLOv8-TO (WORK IN PROGRESS)

Code for the article "From Density to Geometry: YOLOv8 Instance Segmentation for Reverse Engineering of Optimized Structures"

Overview
Reference
Installation
- Prerequisites
- Installing
Datasets

Overview

Brief description of what the project does and the problem it solves. Include a link or reference to the original article that inspired or is associated with this implementation.

Reference

This code aims to reproduce the results presented in the research article:

@misc{rochefortbeaudoin2024density,
      title={From Density to Geometry: YOLOv8 Instance Segmentation for Reverse Engineering of Optimized Structures}, 
      author={Thomas Rochefort-Beaudoin and Aurelian Vadean and Sofiane Achiche and Niels Aage},
      year={2024},
      eprint={2404.18763},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

Installation

Prerequisites

This package comes with a fork of the ultralytics package in the yolov8-to directory. The fork is necessary to add the functionality of the design variables regression.

Installing

git clone https://github.com/COSIM-Lab/YOLOv8-TO.git
cd YOLOv8-TO
pip install -e .

Datasets

Links to the dataset on HuggingFace:

YOLOv8-TO_Data

The Huggingface dataset contains the following datasets (see paper for details):

MMC
MMC-random
SIMP
SIMP_5%
OOD

If you want to use one of the linked datasets, please unzip it inside of the datasets folder. Training labels are provided for the MMC and MMC-random data. To train on the data, please update the data.yaml file with the correct path to the dataset.

path:  # dataset root dir

Spaces:

tomrb
/

YOLOv8-TO

Sleeping