Embeded fusion of visual and acoustic for active acoustic source detection with SGGMM

In this paper, we investigate the problem of reliable detection and localization of active sound source using a new fusion approach of the vision and the acoustic data. The usefulness of the solution is fundamental for both video surveillance and video conference systems. In this aim, we propose combining the two heterogeneous modalities of data by augmenting the 3-D vector of RGB colors used by the Spatially Global Gaussians Mixture Model (SGGMM) for background modeling and segmentation using the acoustic Data. The proposed model provides accurate detection of the targets of interest and evaluation results using an implementation version on wireless sensors network (WSN) of the fusion approach shows performance improvement of the proposed detection and localization solution. This technique enabled a better detection of the moving acoustic source in comparison with the SGGMM only.