PSO Tuned Neural Network for False Contour Reduction

Main Article Content

Regina Manicka Rajam.G


In this project, the adaptation of network weights using Particle Swarm Optimization (PSO) was proposed as a mechanism to improve
the performance of Artificial Neural Network (ANN) in modelling a image false contour reduction Technique. The false contour reduction part
has already two steps. They are NN processing and bi-directional filtering. False contours are reduced by pixel wise processing using NNs in the
first step and bi-directional smoothing is applied to a neighbouring region of the false contour in the second step. PSO is proposed to allow
automatic update of network weights to increase the adaptability to dynamic environment. The results obtained in this paper confirmed the
potential of PSO-based ANN model to successfully model false contour reduction process. The results are explored with a discussion using the
SNR illustrate the usability of the proposed approach. Finally, conclusions and future works are derived.



Keywords: Bi-directional smoothing, false contour reduction, decontouring, false contour detection, neural network


Download data is not yet available.

Article Details