Asian Journal of Engineering and Technology Innovation

Survey Paper on Vector valued image processing using parallel level sets

Chitrashi Singh, Deep Shankar Roy,Ramendra Singh Rathore, Roshani Debnath

Computer Engineering Department, Dr. D Y Patil College of engineering,
Pune University, Maharashtra, India

Abstract

The use of digital image processing techniques has many advantages in terms of enhancing the quality of a picture. Most of the image processing tools use the CMYK model but the RGB color model yields a better output. There is a strong inter-channel correlation in vector valued images which is yet to be explored. Based on the angle between the spatial gradients of their channels, we put forward a new perception of treating vector-valued images. We can attain our desired objective, minimizing the cost functional which reprimands large angles. Placing forward the idea and the corresponding cost functional we discuss their Gâteaux derivatives that lead to a diffusion-like gradient descent scheme. Denoising and demosacking help in exemplifying the properties of cost functional. For low level noised images demosacking will provide superior results.

Keywords:

Parallel level sets, vector-valued images, variational methods, non-linear diffusion, denoising, demosaicking.