Skip to content

Latest commit

 

History

History
12 lines (10 loc) · 687 Bytes

File metadata and controls

12 lines (10 loc) · 687 Bytes

A two step unsupervised FBH detection in Guided Wavefield using K-means and spectral CNN

Two Step-Unsupervised-FBH-detection-Usingand-k-means-and-Spectral-CNN In this reposity, first a train set is obtained with the application of k-means clustering method optimized via Elbow method on a CFRP sample containing several FBHs. The result of clustering provided some TP and FP scan points, constituting the label of the train set. Afterwards, a simple specteral CNN structure is used to construct the final map.


Flowchart