Vinay, A Vikas, S A. creator text xx Bengaluru Jan-May 2019 PES University 9999 monographic und Eng lis h

print

The music and speech discrimination in an audio signal have many practical applications in the multimedia domain such as automated speech recognition and classification, radio broadcast monitoring. This task involves segmentation of audio samples and classification of each sample into speech or music. The objective of the current work is to investigate implementation of audio segmentation for speech and music discrimination using SNN (Spiking Neural Network). The current approach uses CNN (Convolutional Neural Network) with SNN for recognition and classification of music and speech bits in an audio file through spectrograms. The Spiking Neural Networks are a class of ANNs (Artificial Neural Network) that more closely resemble natural and biological neural network that work on STDP (Spike Timing Dependent Plasticity) rule. The proposed approach focuses on implementation of a model to process the dataset that comprises of audio files covering wide range of music and speech sam 231221