Publications

Abstract : During denoise an image; noise level estimation is one of the most important key factors. The accurate noise level estimation is needed before processing the image. The prior knowledge of noise level estimation is also used for restoring the image without degradation. In this proposed work, the noise level is estimated by observed singular values on noisy images. The proposed work has two new methods for addressing the main challenges of the noise level estimation.1.The tail magnitude value of the noisy images singular values has high compare with signal image. This aspect is used for estimate the noise level. 2. The visual based Gaussian noise estimation is used for preprocessing the many 2D- signals processing application which enhance the range of this work. The experimental result for this noise level estimation provides reliable and also applicable for real time images/frames and some special images such as cartoon. The proposed work is needed a simple processing unit for implementing in hardware and results are more accurate. It can be used to pre-processing all kinds of real time images.