SBIR-STTR Award

We propose a new approach to automatic target recognition based on the concept of stochastically deformable templates. In contrast to conventional static approaches, our concept is based on active image tracking. It allows for simultaneous image processing, parameter estimation, and filtering. High target signature variability captured by optical, infrared, millimeter-wave and microwave radar acoustic sensors, etc. substantially diminishes the effectiveness of these technologies in automatic target recognition systems. In this proposal, we address related problems of effective representation and processing of target signatures subject to random fluctuations. Our approach is to represent randomly perturbed dynamically evolving images using the emerging mathematical technique of wavelet descriptors for stochastically deformable templates. The wavelet descriptor has many desirable properties including multiresolution representation, invariance, uniqueness, stability and spatial localization. Real time processing of randomly evolving target signatures is seriously restricted by limitations on availability of communication channels, bandwidth length, computer power, etc. One way to circumvent these difficulties is the dynamic distribution of resolution, so that the maximum resolution windows are applied to a small number of the image landmarks. To achieve this goal, we propose a new method of template deformation tracking and prediction, based on the nonlinear filtering approach. Specifically, we use a new fast spectral algorithm for nonlinear filtering. By separation of observations and parameters, this algorithm shifts the main bulk of computations off line. The main objective of the project is to demonstrate that the recent breakthroughs in wavelet-based image descriptors and fast nonlinear filtering make feasible the real-time implementation of our approach.

In Phase I we demonstrated the feasibility of a new approach to automatic target recognition and tracking. This approach was made possible by recent advances in fast nonlinear filtering and target signature modeling via stochastic deformable wavelet templates (SDWT). In contrast with conventional approaches to target recognition, our new technology is based on the synergy of active tracking and multiresolution image processing. The resulting methods are applicable to the identification and tracking of agile targets. In the basic Phase II effort, we will improve on this technology, implement it in the form of prototype software, and perform preliminary tests on real as well as simulated data. The main components of this software package include: databases for milifacet IR and optical signatures; software for nonlinear filtering, prediction and smoothing; and software for target identification based on the fusion of radar tracking and imaging. We plan to apply these software tools to target recognition, and expecct to demonstrate a 100 to 1 improvement in recognition time compared to that attained using traditional methods. In addition, we investigate techniques to increase recognition accuracy, e.g. image and video enhancement via IR sensor noise modeling, wavelet denoising, and motion compensation. One product of this phase will be SDWT based script and signature recognition software to support high speed optical recognition scanners. In option Phase II our plan is to enhance, finetune and fully test the prototype software developed in basic Phase II. The pimary goal here is to integrate this software with current Navy applications and benchmark the resulting architecture. This phase will also provide for the transition to commercial development (Phase III).