Recognition and Extraction of Required Signals Of Interest in Audio Signals


Authors : Pavan Kumar; Shreya K Shetty; Prajwal Diwakar; Swasthik Narayan Bhat; Dony D’souza

Volume/Issue : Volume 7 - 2022, Issue 6 - June

Google Scholar : https://bit.ly/3IIfn9N

Scribd : https://bit.ly/3nRH1t2

DOI : https://doi.org/10.5281/zenodo.6814658

Without knowing all of the mixing matrix and source signal properties, blind source separation is a process of differentiating sources of signals mixed with time, location, and frequency. This algorithm is widely chosen since it not only has a fast convergence time but also performs well in separation. Blind Source Separation (BSS) is a method of isolating sources from a combination without first determining which sources are present. To accomplish the above objective, several methods have been offered, one of which is ICA (independent component analysis). Duet presupposes that signals are non-overlapping in the temporal domain, while independent component analysis asserts that signals are linearly separable. In the event of blind source separation, ICA is used. They are quite similar to the older traditional fast fixed- point (FastICA) methods based on kurtosis, but they differ in that they are more efficient in terms of processing performance than the corresponding latter ones, which is especially noticeable when the number of samples is large. The Cocktail Party Problem degrades the quality of discourse in this situation. Cocktail party issues are defined as a mix of several sources of speech signals picked up by a microphone. ICA, which has the ability to split many voice signals into individual ones, can be used to solve the aforesaid problem. Gradient and FastICA algorithms are used to apply the principle of negentropy by maximizing non- Gaussianity approaches in ICA. The MATLAB results show that Fast ICA has a faster execution time than gradient with the small number of iterations. Independent component analysis was used to estimate source signals using mutual information minimization. Maximizing non-Gaussianity improves performance in certain cases. The use of kurtosis and negentropy are two strategies for maximizing non-Gaussianity. Because kurtosis is the most vulnerable to outliers and is a computationally resilient process, negentropy is more dependable. For syntactic verification of mechanical and structural systems, the two routine BSS solutions are ICA and SOBI. The traditional FastICA methods are still the most widely used tools for estimating independent components. FastICA functions can be used in a variety of computer languages, including MATLAB.

Keywords : Blind Source Seperation, FastICA Algorithm, PCA Algorithm, Second Order Blind Identification, Negentropy, Kurtosis.

CALL FOR PAPERS


Paper Submission Last Date
31 - March - 2024

Paper Review Notification
In 1-2 Days

Paper Publishing
In 2-3 Days

Video Explanation for Published paper

Never miss an update from Papermashup

Get notified about the latest tutorials and downloads.

Subscribe by Email

Get alerts directly into your inbox after each post and stay updated.
Subscribe
OR

Subscribe by RSS

Add our RSS to your feedreader to get regular updates from us.
Subscribe