Title: Evolutionary neural network for noise cancellation in image data

Authors: Prajna Parimita Dash; Dipti Patra

Addresses: Department of Electrical Engineering, National Institute of Technology, Rourkela, 769008, Orissa, India. ' Department of Electrical Engineering, National Institute of Technology, Rourkela, 769008, Orissa, India

Abstract: A novel artificial neural network called as modified functional link artificial neural network has been proposed for denoising of digital image corrupted with additive white Gaussian noise. Some of the variants of neural network like multilayer perceptron (MLP), direct linear artificial feed-through neural network (DLFANN), functional link artificial neural network (FLANN) has already been implemented in this regard. In FLANN and M-FLANN structure, some of the expanded inputs are chosen using an evolutionary technique (genetic algorithm), called as evolves inputs, to achieve the desired model. These two structures are also trained by genetic algorithm. The potential of the proposed method has been assessed and compared with the existing algorithms. The results showed the superior performance of the proposed method over its counterparts.

Keywords: direct linear feed-through ANNs; artificial neural networks; DLFANN; functional link ANNs; FLANN; Gaussian noise; genetic algorithms; noise cancellation; image data; digital images; image corruption.

DOI: 10.1504/IJCVR.2011.042839

International Journal of Computational Vision and Robotics, 2011 Vol.2 No.3, pp.206 - 217

Published online: 06 Oct 2011 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article