Source Separation with Character Matching
نویسندگان
چکیده
The cocktail party effect describes human's ability to follow one voice source from a mixture of conversations, and often with the addition of background noises. These conversations may happen simultaneously. However, it is tricky for computers to handle this sort of auditory source separation problems. One relatively successful approach is to use Independent Component Analysis (ICA) to separate voice sources from a mixture of them. Specifically in this project, FastICA, an efficient and popular algorithm for ICA is implemented. We then apply a correlation filter called the Minimum Average Correlation Energy (MACE) filter to match the separated voices with their characters for the purpose of identification. Both the methods have been validated by experiments.
منابع مشابه
Estimation of Direction of Arrival Using Matching Pursuit and Its Application to Source Separation
In this paper, we describe a new blind source separation (BSS) method that uses spatial information derived from the direction of arrival (DOA) estimates of each direct and reflected sound. The method we proposed has the following steps: (1) each DOA is estimated using matching pursuit and re-optimized after each new DOA is estimated, (2) using these DOA estimates, the mixing matrix is also est...
متن کاملNon-orthogonal Frequency Hopping Signal Underdetermined Blind Source Separation in Time-Frequency Domain
In this paper, a novel Matching Optimization Algorithm (MOA-algorithm) based on underdetermined blind source separation is proposed for non-orthogonal frequency hopping signal (that is, inner products are not always equal to zero in the same time-frequency point). Compared to traditional methods, the separation method is formulated as matching optimization. In our method, we accomplish the unde...
متن کاملمقایسه ابعاد سرشتی- منشی شخصیت بیماران دارای اختلال اضطراب جدایی بزرگسالی با افراد سالم
Objectives: The purpose of this study was to compare temperament-character dimensions of personality among patients with separation anxiety disorder and healthy person. Method: This research has Ex Post Facto Research Design. The sample consisted of 200 female students from Tehran and Shahed universities who were selected through convenient sampling method. The participants were assessed by the...
متن کاملCorrelation matching approach to source separation in the presence of spatially correlated noise
This paper addresses a new method of source separation that is robust in the presence of spatially correlated but temporally white noise. The method exploits the non-vanishing temporal structure of sources to reduce the effect of noise. In the framework of correlation matching, we develop two efficient least squares algorithm. Useful behavior of the proposed algorithms is confirmed by numerical...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2010