A hybrid method based on time–frequency images for classification of alcohol and control EEG signals

摘要: Classification of alcoholic electroencephalogram (EEG) signals is a challenging job in biomedical research for diagnosis and treatment brain diseases people. The aim this study was to introduce robust method that can automatically identify EEG based on time–frequency (T–F) image information as they convey key characteristics signals. In paper, we propose new hybrid classify the control proposed scheme images, texture feature extraction nonnegative least squares classifier (NNLS). T–F analysis, spectrogram short-time Fourier transform considered. obtained images are then converted into 8-bit grayscale images. Co-occurrence histograms oriented gradients (CoHOG) Eig(Hess)-CoHOG features extracted from Finally, fed NNLS input To verify effectiveness approach, replace by artificial neural networks, k-nearest neighbor, linear discriminant analysis support vector machine separately, with same features. Experimental outcomes along comparative evaluations state-of-the-art algorithms manifest outperforms competing algorithms. experimental promising, it be anticipated upon its implementation clinical practice, will alleviate onus physicians expedite neurological research.