Ellipsoidal Basis Functions for Acoustic and Electro-Optic Data Fusion

The objective of this proposal is to correlate and fuse data from electro-optic and AQS20(A) acoustic sensor data and provide a single computer-aided mine identification system thus significantly increasing mine identification capability by utilizing thecomplementary strengths of both types of sensors. The approach will use 2 levels of fusion. The first level (Lower level fusion) will fuse feature vectors, using the ellipsoidal basis functions (EBF) method, obtained from the two sensor types and provide asingle identification, thus utilizing greater information afforded by two sensors as opposed to using single sensor data. The second level (Higher level fusion) will consider the fusion of a candidate complementary algorithm in addition to the EBF method.The idea here, discussed in the literature, is that each algorithm has its strengths in suppression of specific false alarms and thus a fusion of complementary algorithms will provide optimal performance. The proposed investigation will provide valuabletools and validation for applications of the Ellipsoidal Basis Functions method to areas other than mine-hunting also such as missile and homeland defense as well as object and face recognition technologies. There currently exist various face recognitiontechnologies that have made way into some applications (e.g. Connecticut DMV, West Virginia Child ID programs). The deployed software is still being tested and refined. Most of the techniques currently use a principal components analysis (PCA) followed byextraction of