What question did this study set out to answer?

To improve segmentation precision of aero-engine defects, which are characterized by micro-scale and high-frequency details.

May 6, 2026ComputationOpen Access

A Spectrum-Driven Hierarchical Learning Network for Aero-Engine Defect Segmentation

View Full Paper

Ask AI

Bookmark

Authors

This approach enhances defect segmentation accuracy in aero-engines, indicating improvements in boundary localization and robustness.

To improve segmentation precision of aero-engine defects, which are characterized by micro-scale and high-frequency details.
Developed a spectrum-driven hierarchical learning network for defect segmentation.
Constructed a dual-band spectral module to separate high-frequency and low-frequency components.
Designed a detail-guided module for high-frequency features to improve boundary recovery.
Achieved mIoU scores of 89.82% and 91.44% on Turbo19 and NEU-Seg datasets respectively.
Improved over the second-best method by 5.22% and 4.42% in accuracy.

Xie et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark