I-MATH Associates, Inc. and the NYU Courant Institute of Mathematical Sciences propose the development of algorithms and real time software for the flision of 3D (and higher order) imagery and information. The fundamental technique we will use is geometric hashing, originally conceived by NYU and subsequently applied by I-MATH to military automatic target recognition applications. Hashing is an efficient method for storing a very large quantity (up to tens of 1000's) of models representing various object types and poses, and then quickly deternning which model best represents an unknown item, whose corresponding features are sifted through the hash table. The overall objective of this STTR is not just multidimensional pattern recognition, but rather the maximum extraction of information from multiple sources, which may be dissimilar and perhaps not even imaging. Hashing directly supports such fusion, since multiple types of features can be the basis for an nD hash table. In fact, the features do not even have to have the same dimensionality, the formulation for which has been recently developed at NYU. Phase I will produce two algorithms/software/CPU prototypes, both performing 3D hashing in a 3D space. One prototype will be hosted on an NYU Reterogeneous Farm of RISC and/or SUN computers; this configuration will be focused towards a general computer lab environment. The second prototype will be hosted on a TI Multimedia Video Processing CPU, the TMS 320C8X; this prototype will be focused towards tactical military applications, such as Ladar ATR. Phase II will complete the development of the two prototypes and extend their capability into a higher order dimensionality/space.
Keywords: geometric hashing weighted voting information fusion bayesian reasoning
