Defect Detection in Tire X-Ray Images: Conventional Methods Meet Deep Structures

arXiv:2402.18527v1 Announce Type: cross
Abstract: This paper introduces a robust approach for automated defect detection in tire X-ray images by harnessing traditional feature extraction methods such as Local Binary Pattern (LBP) and Gray Level Co-Occurrence Matrix (GLCM) features, as well as Fourier and Wavelet-based features, complemented by advanced machine learning techniques. Recognizing the challenges inherent in the complex patterns and textures of tire X-ray images, the study emphasizes the significance of feature engineering to enhance the performance of defect detection …

abstract advanced arxiv automated binary cs.cv cs.lg defect detection detection eess.iv extraction feature feature extraction features fourier images machine matrix paper ray robust type wavelet x-ray