Sound Classification using Image Feature Extraction Technique

Abstract

This paper presents texture feature extraction methods for sound classification. Nowadays, many researchers interest in combination of digital signal processing and digital image processing fields to get higher efficiency. In this paper, feature extraction methods used in image processing are applied in classification of signal processing. Signals are converted into image format and then features are extracted using bi-directional local binary pattern. Feature vector is constructed using these features and then label the input signal by checking similarly value from known dataset using multi support vector machine classifier. Evaluation is tested on benchmark dataset namely ESC10 Dataset, ESC50 Dataset and UrbanSound8K Dataset. This paper presents texture feature extraction methods for sound classification. Nowadays, many researchers interest in combination of digital signal processing and digital image processing fields to get higher efficiency. In this paper, feature extraction methods used in image processing are applied in classification of signal processing. Signals are converted into image format and then features are extracted using bi-directional local binary pattern. Feature vector is constructed using these features and then label the input signal by checking similarly value from known dataset using multi support vector machine classifier. Evaluation is tested on benchmark dataset namely ESC10 Dataset, ESC50 Dataset and UrbanSound8K Dataset.