Playback detection using machine learning with spectrogram features approach

Fault Diagnosis of High-speed Train Bogie Based on Spectrogram and Multi-channel Voting

“…M Mustafa et al extracted Gray Level Cooccurence Matrix texture features from Electroencephalogram spectrograms and used KNN for classification [7]. J Dembski et al used histograms of Oriented Gradients (HOG) and SVM to playback detection in automatic speaker verification systems [8]. In this paper, the data used was acquired from a software named SIMPACK with a certain train model and the track spectrum of Wuhan-Guangzhou line.…”