Optimal data fusion of correlated local decisions in multiple sensor detection systems

Information assimilation framework for event detection in multimedia surveillance systems

“…For example, featurelevel (early) fusion of video and audio has been proposed for the problems speech processing [Hershey et al 2004] and recognition [Nefian et al 2002], tracking [Checka et al 2004], and monologue detection [Nock et al 2002] by using the mutual information among the video and audio features under the assumption that audio and video signals are individually and jointly Gaussian random variables. On the other hand, late fusion strategies have also been used in sensor fusion applications [Rao and Whyte 1993], [Chair and Varshney 1986], [Kam et al 1992]. In late fusion strategy, a global decision is made by fusing the local decisions obtained from each data source.…”

“…[Chair and Varshney 1986] established an optimal fusion rule with the assumption that each local sensor made a predetermined decision and each observation was independent. [Kam et al 1992] generalizes their solution for fusing the correlated local decisions.…”

Distributed inference in wireless sensor networks

“…The findings in this paper indicate that, even in this simple setting, an optimal detector may exhibit very complicated behaviour. In Kam et al [37], the structure of an optimal fusion rule has been examined for the more encompassing scenario where multiple binary sensors observe conditionally dependent random variables. In Blum & Kassam [38], the structure of an optimal detector when faced with weak signals and dependent observations has been explored.…”

Mathematical Basis of Sensor Fusion in Intrusion Detection Systems

“…Drakopoulos & Lee (1995) derive an optimum fusion rule for the Neyman-Pearson criterion, and uses simulation to study its performance for a specific type of correlation matrix. Kam et al (1995) considers the case in which the class-conditioned sensor-to-sensor correlation coefficient are known, and expresses the result in compact form. Their approach is a generalization of the method adopted by Chair & Varshney (1986) for solving the data fusion problem for fixed binary local detectors with statistically independent decisions.…”

“…Their approach is a generalization of the method adopted by Chair & Varshney (1986) for solving the data fusion problem for fixed binary local detectors with statistically independent decisions. Kam et al (1995) uses Bahadur-Lazarsfeld expansion of the probability density functions. Blum et al (1995) study the problem of locally most powerful detection for correlated local decisions.…”