Reform and practice of a ceramics-focused interdisciplinary undergraduate teaching model in new engineering education—An innovative teaching case study on defect detection in ceramic bearings

Implementing the new engineering education philosophy of “promoting learning through application and enhancing teaching through research,” advancing teaching reforms in interdisciplinary courses, this paper employs ceramic bearing defect detection as a case study to explore the design and implementation of teaching case studies to promote the deep integration of theory, technology, and application. The course focuses on detecting surface defects in ceramic bearings. Defects exhibit characteristics of minute dimensions, complex spatial distribution, and uneven gray-scale distribution. Thus, a context-aware feature pyramid network with attention guidance is constructed to achieve multi-scale feature fusion. The detection capability of minute defects is enhanced by integrating deep semantic information with shallow localization features. A genetic clustering module is designed to adaptively fit annotation distributions and optimize candidate box dimensions, thereby further enhancing detection accuracy and efficiency. Teaching practice demonstrates that this interdisciplinary integration approach effectively enhances the performance and engineering applicability of defect detection. Students’ comprehension and application of theoretical knowledge are also improved. It offers a practical model for advancing teaching and research synergistically in the innovation of high-end equipment nondestructive testing technologies, aligned with the goals of new engineering education.