Abnormal acoustic event localization based on selective frequency bin in high noise environment for audio surveillance

Abstract

In this paper, a method for source localization for surveillance system is presented. In particular, we propose an algorithm for abnormal acoustic event localization based on a novel approach of relevant frequency bin selections by statistical analyses. By means of selective frequency bin, it becomes possible to localize the event more accurately in high noise environment with low computational complexity. The effectiveness is verified through the experimental results in varied noise environments with different levels of Signal to Noise Ratio (SNR).

Generalized cross-correlation based noise robust abnormal acoustic event localization utilizing non-negative matrix factorization

“…An overview of the propose Figure 1 [3]. Pre-trained selected ach abnormal acoustic event e weights being fixed, there noise levels present in the cting (or weighting high) n non-stationary noise, we nction using Non-negative he first step of the proposed ormal event bases and their training with labeled set of .…”

“…By assuming we have equation (4) where sired signal and noise. (4) s to factorizing equation (3). n distance measure which is f D, H is as follows.…”

“…Finding abnormal acoustic event loca active research area with applicatio systems [1][2][3] . In recent years, it becam in surveillance systems related to acous as sudden noise detection and localizatio scream detection [1][2].…”