Informative Acoustic Feature Selection to Maximize Mutual Information for Collecting Target Sources
Abstract: An informative acoustic-feature feature-selection selection method for collecting target sources in noisy environments is proposed. Wiener filtering is a powerful framework for sound-source source enhancement. For Wiener Wiener-filter filter estimation, statistical-mapping statistical functions, such as deep ep neural network based or Gaussian mixture model based mappings, have been used. In this framework, it is essential to find informative acoustic features that provide effective cues for Wiener Wiener-filter filter estimation. In this study, we measured the informativen informativeness ess of acoustic features using mutual information between acoustic features and supervised Wiener Wiener-filter filter parameters, e.g., prior signal-to-noise noise ratios, and developed a method for automatically selecting informative acoustic features from a large number of feature candidates. To automatically select optimum features, we derived a differentiable objective function in proportion to mutual information based on the kernel method. Since the higher order correlations between acoustic features and Wiener-filter parameters rameters are calculated using the kernel method, the statistical dependence of these variables is accurately calculated; thus, only meaningful acoustic features are selected. Through several experiments conducted on a mock sports field, we confirmed that tthe signal-to-distortion distortion ratio score improved when various types of target sources were surrounded by loud cheering noise.