A novel method for automatic image registration based on wavelet and near fuzzy set

Abstract

Automatic image registration is still a major challenge in many of the image processing applications, to name a few-remote sensing, medical imaging, industrial image analysis etc. In general, the problem of image registration can be identified as the determination of translations and a small rotation between the respective source images and generation of the resulting registered images. The most critical issue in regards to appropriate image registration is the variability in terms of the different image sensors in producing the source image, which can affect the accuracy in the resultant registered image. In this paper, we have proposed a novel image registration technique based on wavelet theory and near-fuzzy set approach. We have used five sets of test images for our experiment and the experimental results for the entire test sets are superior in terms of noise reduction and varied difference in the image content compared to the other related research works. To the best of our knowledge, our approach of image registration using near-fuzzy set approach is first of its kind and the superior quality of the resultant registered image can well justify its novelty.