🤖 AI Summary
This study addresses the lack of interpretability and decision traceability in deep learning models for detecting defects in aerospace-grade SiC/SiC composites using X-ray computed tomography (XCT). To this end, the authors propose p-ResNet-50, a framework that integrates a prototype layer into ResNet-50 to explicitly link classification outcomes to expert-defined semantic defect categories. Anchor and medoid regularization are introduced to mitigate prototype collapse, while UMAP-based latent space visualization enables clear identification of regions with high prediction uncertainty. Evaluated on approximately 12,000 XCT image patches, the method achieves an accuracy of 0.957 and a ROC-AUC of 0.994—matching the performance of black-box baselines—while simultaneously delivering auditable and human-interpretable decision rationales.
📝 Abstract
Non-destructive testing of aerospace SiC/SiC composites via X-ray computed tomography (XCT) relies on expert visual assessment, with current workflows offering limited traceability for accept/reject decisions. Deep convolutional networks can automate defect detection, yet their black-box nature conflicts with the transparency that industrial inspection practice demands. To close this gap, we introduce p-ResNet-50, a convolutional framework extended with a prototype layer that couples high detection accuracy with case-based explanations. Six learned prototypes are explicitly aligned with expert-defined semantic categories-healthy matrix, matrix--air interfaces, pores, line-like defects, and mixed morphologies-so that every classification is traceable to a physically meaningful reference. Two novel regularisation terms, anchor-based and medoid-based, tether prototypes to expert-selected patches and prevent prototype collapse, addressing a known limitation of prototype networks. Latent-space analysis via UMAP delineates semantically coherent sub-domains and maps zones of uncertainty where misclassifications concentrate, giving inspectors an explicit picture of where the model is-and is not-reliable. The framework is validated on an XCT patch dataset of approximately 12,000 patches extracted from four defect-rich SiC/SiC laboratory specimens. Taking a black-box ResNet-50 as a baseline (ROC-AUC = 0.991), the prototype extension achieves comparable performance (accuracy 0.957 vs. 0.959; ROC-AUC 0.994 vs. 0.993) while trading a slight reduction in sensitivity for higher precision and specificity. Each decision is backed by representative evidence patches, and the model explicitly flags its uncertainty regions. Beyond defect mapping, the framework establishes a reusable methodology for embedding domain-expert knowledge into prototype networks, applicable to other XCT inspection scenarios requiring traceable, auditable decisions.