Predicting Motorcycle Tire Failure with Deep Learning

Motorcycle tire maintenance is crucial for preventing accidents in the Philippines, where motorcycles are a primary mode of transportation. This study proposes a deep learning-based model to assess motorcycle tire condition and predict the need for replacement. The model utilizes Mask R-CNN for tire segmentation and a Convolutional Neural Network (CNN) for image classification. Preprocessing techniques, including normalization, grayscale conversion, and histogram equalization, enhance image quality. Trained on a dataset of 370 motorcycle tire images, the model achieved an accuracy of 95.9% in distinguishing between safe and worn tires. This research holds significant potential for improving road safety by providing an automated tool for tire condition assessment, ultimately reducing accidents caused by tire failure. Furthermore, the model can be integrated into mobile applications or motorcycle maintenance services, empowering riders and mechanics to make informed decisions about tire care.