Distributed robust image mosaics

Abstract

In this paper, we describe a distributed robust image mosaics system that is robust to outliers and is better in computation efficiency. The system integrates a Web-based user interface, robust image mosaics, and a distributed computing model. In the robust image mosaics, we take a feature-based approach for image registration, and a skipped mean estimator is incorporated into the Levenberg-Marquardt method for robust nonlinear parameter estimation. The distributed computing model is used to speed up the computation of image mosaics on a network of workstations. Load balancing strategies are proposed to improve the efficiency of the parallel computation under uneven loads. Experimental results show that our system can suppress mismatching during image registration and provide significant speedup in computation.