A novel framework for under-determined blind source separation based on adaptive source counting using mixed linear and circular data clustering algorithm for low latency applications
Multimedia Tools and Applications, ISSN: 1573-7721
2024
Metric Options: CountsSelecting the 1-year or 3-year option will change the metrics count to percentiles, illustrating how an article or review compares to other articles or reviews within the selected time period in the same journal. Selecting the 1-year option compares the metrics against other articles/reviews that were also published in the same calendar year. Selecting the 3-year option compares the metrics against other articles/reviews that were also published in the same calendar year plus the two years prior.
Example: if you select the 1-year option for an article published in 2019 and a metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019. If you select the 3-year option for the same article published in 2019 and the metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019, 2018 and 2017.
Citation Benchmarking is provided by Scopus and SciVal and is different from the metrics context provided by PlumX Metrics.
Example: if you select the 1-year option for an article published in 2019 and a metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019. If you select the 3-year option for the same article published in 2019 and the metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019, 2018 and 2017.
Citation Benchmarking is provided by Scopus and SciVal and is different from the metrics context provided by PlumX Metrics.
Article Description
Underdetermined Blind Source Separation (UBSS) refers to a general class of signal processing algorithms, aiming to recover the underlying source signals from related mixtures without resorting to any prior information about the mixing matrix system and with less sensors than source signals. Technologies such as teleconferencing, hands-free telephony and hearing aids mostly require real-time processing. This matter is a major challenge in BSS problem, as traditional methods generally require significant amounts of data to generate sufficient statistics for separation. So, the proposed method is trying to give an overall solution for low latency applications. For this purpose, the High-Resolution Sub-Band Decomposition (HRSBD) algorithm in sparse time domain is utilized to compensate the low data efficiency of short time block lengths. Additionally, in contrast to the conventional methods which presume the known number of source signals during processing frames, we derive an adaptive source counting procedure in each processing block. Some clustering techniques assume a linear distribution for the audio mixture, whereas some others utilize circular statistics. Most of the linear approaches lead to failure in the case of directional data and most of the circular methods are unable to generalize in the case of multi-dimensional setup (multiple-microphone setup). Our research proves that the nature of mixing data in UBSS problem has both linear and circular features. Therefore, a new clustering scheme is proposed which effectively addresses the directional data and multi-dimensional setup simultaneously. Finally, we propose a separation procedure with low computational complexity which selects the best separation method considering the results of adaptive source counting phase of the algorithm. Experimental evaluations show the superiority of the proposed method against the state-of-the-art techniques according to the remarkable and dominant high performance of estimating the number of sources, mixing matrix estimation accuracy and the effective results of the source separation in both instantaneous and reverberant conditions.
Bibliographic Details
Springer Science and Business Media LLC
Provide Feedback
Have ideas for a new metric? Would you like to see something else here?Let us know