Combined curvelet and ASF with neural network for denoising sonar images

Curvelet transform is the member of multiscale geometric transforms, which provides an best solution to the problems associated with image denoising with wavelets. The wavelet´s performance degrades when circular type space-frequency like curve or arc edges or lines are viewed in the image. Curvelet transform has redundant dictionary that can provide sparse representation of signals that have edges along regular curve. The second generation curvelet transform is simpler to understand and use and also faster and less redundant compared to its previous version. Curvelet transform is depicted in various domains and meant for higher dimensions. This paper focused on an better curvelet based denoising technique to improve sonar images.