Research Article
Schizophrenia Identification Through Deep Learning on Spectrogram Images
@INPROCEEDINGS{10.1007/978-3-031-48888-7_1, author={Amarana Prabhakara Rao and G. Prasanna Kumar and Rakesh Ranjan and M. Venkata Subba Rao and M. Srinivasulu and E. Sravya}, title={Schizophrenia Identification Through Deep Learning on Spectrogram Images}, proceedings={Cognitive Computing and Cyber Physical Systems. 4th EAI International Conference, IC4S 2023, Bhimavaram, Andhra Pradesh, India, August 4-6, 2023, Proceedings, Part I}, proceedings_a={IC4S}, year={2024}, month={1}, keywords={Schizophrenia Electroencephalogram Spectrogram Short-time Fourier Transform Convolutional Neural Networks Deep Learning Approaches}, doi={10.1007/978-3-031-48888-7_1} }- Amarana Prabhakara Rao
G. Prasanna Kumar
Rakesh Ranjan
M. Venkata Subba Rao
M. Srinivasulu
E. Sravya
Year: 2024
Schizophrenia Identification Through Deep Learning on Spectrogram Images
IC4S
Springer
DOI: 10.1007/978-3-031-48888-7_1
Abstract
Schizophrenia (SZ) is one of the mental disorder due to which many people are suffering around the world. People suffering with this disorder experience hallucinations, delusions, confusing speech and thinking patterns, etc. In a clinical environment, doctors judge Schizophrenia directly using electroencephalogram (EEG). Automatic detection of SZ is achieved in earlier works by using the time domain and frequency domain features extracted from the given EEG signals. These features are used to train various Machine Learning and Deep Learning approaches for the classification of SZ from the given EEG signal. The proposed work uses Short-Time Fourier Transform (STFT) for converting 1D EEG data into 2D spectrogram image data. This work proposes a simple Convolutional Neural Network (CNN) model for the efficient detection of SZ from the given spectrograms. Performance of the proposed CNN model is compared with various existing CNNs such as Alex net, VGG16, Resnet. Performance of these CNNs is evaluated in terms of accuracy, precision, recall and F1 Score. It is observed from the results that the proposed CNN performed better showing its potential for efficient detection of SZ.
