August 15, 2025SensorsOpen Access

Bridge Damage Identification Using Time-Varying Filtering-Based Empirical Mode Decomposition and Pre-Trained Convolutional Neural Networks

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Framework integrates time-varying filtering with convolutional neural networks to improve bridge damage identification accuracy.

Proposed method enhances frequency-domain features while suppressing non-stationary noise through adaptive denoising.
Numerical simulations demonstrate that time-varying filtering-based empirical mode decomposition improves modal purity and frequency separation.
Comparison of pre-trained convolutional neural networks shows ResNet-50 delivers optimal performance in damage classification tasks.
Principal component analysis indicates significant improvements in clustering and feature data separability with better class boundaries.

Zeng et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark