Overview

The main goal of this project is to improve the performance of Automatic Target Recognition (ATR) by developing a more powerful ATR frame work which can handle changes in the appearance of a target more efficiently and robustly. The new framework will be built around a hybrid model of appearance by integrating (1) Algebraic Functions of Views (AFoVs), a powerful mathematical model of geometric appearance, with (2) eigenspace representations, a well known empirical model of appearance which has demonstrated significant capabilities in recognizing complex objects under no occlusion. The hybrid model has significant advantages over geometric and empirical models alone, providing a more realistic model for predicting object appearance. Also it can support various type of data including visual and Synthetic Aperture Radar (SAR) images. To address the requirements of realistic ATR applications, the new framework will be augmented with grouping, indexing, probabilistic hypothesis generation and incremental learning strategies.

This research addresses a problem of fundamental importance to the Office of Naval Research (ONR) and other defense agencies. Potential ONR payoffs include more robust, efficient and general algorithms for target recognition and tracking. Technological gains in developing more effective ATR systems will advance our knowledge in object recognition and will have important and immediate implications to a wide variety of fields currently pursuing object recognition applications such as mobile robot localization, autonomous navigation, monitoring, and surveillance. To ensure the applicability of our results in other related areas, the proposed research will be conducted in close cooperation with ONR, Los Alamos National Laboratory (LANL), and Honeywell. This project is a joint effort between the Departments of Computer Science at the University of Nevada, Reno (UNR) and Department of Mechanical Engineering at the University of Nevada, Las Vegas (UNLV) .