BoltVision is a comparative analysis of various machine learning models including Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), and Compact Convolutional Transformers (CCTs), aimed at accurately classifying missing bolts in train components. This repository contains the code and models used in the study published in Machines (MDPI). For more information, read our paper: BoltVision: A Comparative Analysis of CNN, CCT, and ViT.
We would need pytorch mainly for this, we also have some model with tensorflow
Different models are in different files to test properly. The file names are a good indication.
Not including the dataset and university has the rights to it
This study illustrates the superior assessment capabilities of these models and discusses their effectiveness in addressing the prevalent issue of edge devices. Results show that BoltVision, utilising a pre-trained ViT base, achieves a remarkable 93% accuracy in classifying missing bolts.
If you use BoltVision in your research, please cite our paper:
@article{alif2024boltvision,
title={BoltVision: A Comparative Analysis of CNN, CCT, and ViT in Achieving High Accuracy for Missing Bolt Classification in Train Components},
author={Alif, Mujadded Al Rabbani and Hussain, Muhammad and Tucker, Gareth and Iwnicki, Simon},
journal={Machines},
volume={12},
number={2},
pages={93},
year={2024},
publisher={MDPI}
}