NAECON08 grand challenge entry using the belief filter in audio-video track and ID fusion

Abstract

The IEEE NAECON2008 Challenge Problem competition was to observe, monitor, and determine the cause-effect relations in an audio-video data set. Acoustic-to-visual fusion establishes the relationships of events much as the human brain associates spatial-temporal audio and video feature data. A machine, like a human, can infer from the audio (hear) and the video (see) the cognitive processing relationships between perception and judgment using standard figures of merit (FOM). After downloading the data sets, a series of signal processing, image fusion, and tracking methods were applied to extract salient features for a belief filter (BF) to determine the cause-effect event of ball movement on a suspended object.