Annotated Tyre Image Dataset for Machine Learning-Based Life Prediction
Description
This dataset consists of high-resolution images of vehicle tyres that have been acquired in various states of usage. The images have been gathered and labeled for machine learning and computer vision research in the classification of tyre health and prediction of residual life. Each image has been labeled with relevant metadata, such as usage history when applicable. Researchers and developers can use this dataset to train and develop convolutional neural networks to classify images as either “defective/good,” while also extending models to continuous regression-based prediction tasks as the model predicts remaining life of the tyre.
Files
Steps to reproduce
The dataset was comprised of four car service stations in Meerut to reflect the different tyre condition and environment. Independent images were also captured from a couple of tyre showrooms to take into consideration newer and moderately used tyres. Photos were taken using an android smart phone with a 48 MP triple rear camera for high-resolution and detailed visuals. Each image was thoroughly annotated from experts based on tyre usage. By including tyres from different vehicles and shooting conditions, the dataset reflects real-world diversity reflecting the research problem examined for AI-based tyre classification and residual life predictions.
Institutions
- Dr A P J Abdul Kalam Technical University
- Meerut Institute of Engineering and Technology