No abstract

No abstract

Due to technological improvements many methods have been proposed for speaker verification. While performance is satisfactory with large amounts of speech data, there is significant degradation in performance with short utterances. Many research works have been carried out to handle short utterance issue of the speaker verification systems used in real-world scenario. In this paper we primarily emphasis on the survey of different feature extraction methods for textindependent speaker verification. We first briefly review conventional systems to show its progress. In this work, we present a brief review of features that are used to capture speaker information at different analysis lengths of speech utterance. We also put the major findings and challenges of this research feview in a nutshell.

No abstract

Edges and textures are important features in texture analysis that helps to characterize an image. Thus, the edges and textures must be retained during the process of denoising. In this paper, we present a combined wavelet decomposition and Variational Mode Decomposition (VMD) approach to effectively denoise texture images while preserving the edges and fine-scale textures. The performance of the proposed method is compared with that of wavelet decomposition, VMD and a combined VMD-WT technique. Although VMD-WT outperforms VMD and wavelet decomposition, it is highly dependent on the choice of parameters. The proposed method overcomes the above limitation and also performs better than wavelet decomposition and VMD.