用深度网络提升波场中微裂纹检测精度，解决数据不平衡难题
MicroCrackAttentionNeXt: Advancing Microcrack Detection in Wave Field Analysis Using Deep Neural Networks through Feature Visualization
- 基于波场数据设计新型神经网络，结合特征可视化优化结构
- 在仅5%裂纹像素的极端不平衡数据上实现86.85%检测准确率
- 适合做材料损伤检测、工业无损评估的研究者与工程师参考
利用深度神经网络（DNN）通过波场与损伤区域相互作用的自动化流程进行微裂纹检测备受关注。这些高维时空裂纹数据量有限，且时间维度维度庞大。数据集存在显著类别不平衡，每样本裂纹像素平均仅占总像素的5%。这种极端类别不平衡使深度学习模型在识别不同尺度微裂纹时易受偏差影响，倾向于预测多数类，导致检测精度低下。本研究在先前基准SpAsE-Net（一种非对称编码器-解码器网络）基础上展开，通过曼达拓扑发现与分析（MDA）算法进行特征空间可视化，考察多种激活函数与损失函数的影响。优化后的架构与训练方法最终达到86.85%的准确率。
原文摘要 · Abstract (English)
Micro Crack detection using deep neural networks (DNNs) through an automated pipeline using wave fields interacting with the damaged areas is highly sought after. These high-dimensional spatio-temporal crack data are limited, and these datasets have large dimensions in the temporal domain. The dataset presents a substantial class imbalance, with crack pixels constituting an average of only 5% of the total pixels per sample. This extreme class imbalance poses a challenge for deep learning models with the different micro-scale cracks, as the network can be biased toward predicting the majority class, generally leading to poor detection accuracy. This study builds upon the previous benchmark SpAsE-Net, an asymmetric encoder-decoder network for micro-crack detection. The impact of various activation and loss functions were examined through feature space visualization using the manifold discovery and analysis (MDA) algorithm. The optimized architecture and training methodology achieved an accuracy of 86.85%.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。