Multi-focus color image fusion using NSCT and PCNN

Abstract

Listen

Translate article

Image fusion combines different images of same scene from different sensors or from the same sensor at different times to create a new image. Due to limited depth of focus of optical lens, it is often impossible to acquire an image that contain all relevant focused objects, some objects will be in focus some others will be out of focus. Using multi focus image fusion one can get one image with all of objects in focus. Image fusion methods are usually divided into spatial domain and transform domain techniques. One of the simplest spatial domain methods is block method but it causes block effect. Another important spatial based method is focused region based method which is able to detect the clear regions of source images and then directly copy the pixels from clear region to fused image. However, these methods generate artificial information and discontinuous phenomenon at border of focused region. This will affect visual fidelity of fused image. Compared with spatial based methods, methods using multi scale transform successfully overcome the above mentioned disadvantages. Image fusion methods based on Non-Subsampled Contourlet Transform (NSCT) and perform very well for gray scale images. In this paper, a new method multi-focus image fusion is proposed that is suitable for color images using NSCT and Pulse Coupled Neural Network (PCNN).