Exposure Fusion using Particle Filtering Techniques

Abstract

Fusion of Multiple exposure images has attracted lot of attention over the years. There are various approaches for multi-exposure image fusion, in these approaches the images are treated and fused in the spatial domain or in the transform domain by defining a fusion rule. This filtering based approach for exposure fusion; relies basically on natural spatial, statistics of the images rather than on parameter extraction based approaches. The basic advantage of the particle filtering based approach is its linearity and quick convergence. This method is based on stochastic principles for de-noising. We treat the particle filtering based approach as an image de-noising problem in the spatial domain in this paper.