No abstract

No abstract

This paper presents the automatic genre classification of Indian Tamil music and western music using timbral features and fractional Fourier transform (FrFT) based Mel frequency cepstral coefficient (MFCC) features. The classifier model for the proposed system has been built using K-nearest neighbours and support vector machine (SVM) classifiers. In this work, the performance of various features extracted from music excerpts have been analyzed, to identify the appropriate feature descriptors for the two major genres of Indian Tamil music, namely classical music (Carnatic based devotional hymn compositions) and folk music. The results have shown that the feature combination of spectral roll off, spectral flux, spectral skewness and spectral kurtosis, combined with fractional MFCC features, outperforms all other feature combinations, to yield a higher classification accuracy of 96.05 %, as compared to the accuracy of 84.21 % with conventional MFCC. It has also been observed, that the FrFT based MFCC, with timbral features and SVM, efficiently classifies the two western genres of rock and classical music, from the GTZAN dataset, with fewer features and a higher classification accuracy of 96.25 %, as compared to the classification accuracy of 80 % with conventional MFCC.

T his paper presents th e classification of m usical instrum ents using Mel Frequency C epstral Coefficients (M FCC) and Higher O rder Spectral features. M FCC, cepstral, tem poral, spectral, and tim bral features have been widely used in th e task of m usical in stru m en t classification. As music sound signal is generated using non-linear dynam ics, non-linearity and non-G aussianity of th e musical instrum ents are im p o rtan t features which have not been considered in th e past. In th is paper, hybridisation of M FCC and Higher O rder Spectral (HOS) based features have been used in th e task of musical in stru m en t classification. H O S-based features have been used to provide in stru m en t specific inform ation such as non-G aussianity and non-linearity of th e musical instrum ents. T h e ex tracted features have been presented to C ounter P ropagation N eural Network (CPNN ) to identify th e instrum ents and th eir family. For experim entation, isolated sounds of 19 musical instrum ents have been used from M cGill U niversity M aster Sample (MUMS) sound database. T he proposed features show th e significant im provem ent in th e classification accuracy of th e system. K e y w o rd s : feature extraction; M FCC; HOS; bispectrum ; bicoherence; non-linearity; non-G aussianity; CPNN; Zero Crossing R ate (ZCR).

This paper gives an overview idea of efficient retrieval of images using different Mpeg-7 Features. Content Based Image Retrieval is a technique of automatic indexing and retrieving of images from a large data base. Feature Extraction and Similarity Matching are the two major steps for CBIR Systems. Color, Texture and Shape represent the three visual features for any image. Mpeg-7 Stands for Multimedia Content Description Interface. The main objective of Mpeg-7 is to provide a standardized set of technologies for describing multimedia content. It has allowed quick and efficient content identification, and addressing a large range of applications. The visual descriptors are classified according to the feature such as color, shape, texture, etc. This paper has used Color Structure Descriptor for color and Edge Histogram Descriptor for texture.These two features are also integrated to increase the performance of CBIR Systems. The efficiency of all methods are demonstrated with the help of results.