<![CDATA[Tire damage inspection can be categorized as part of vehicle maintenance with the aim of ensuring that the tire condition is in good condition. Visual inspection using human observation has limitations, making it not always accurate and can result in errors in determining tire suitability. This research designs a machine learning modeling using Convolutional Neural Network (CNN) to detect damage to mobile tires. The parameters used in the CNN model training are the Adam optimizer, learning rate 0.0001, batch size 16, and using the Early Stopping function. In this study, the CNN modeling was tested with two treatments, namely using a dataset without data augmentation and a dataset using data augmentation, then the results were evaluated using a confusion matrix. The results showed that data augmentation treatment can significantly improve model performance, with an increase in accuracy of 20%, precision of 14%, recall of 22%, and f1-score of 19% compared to treatment without data augmentation]]>
Copyrights © 2023
