Optimal Image Restoration Using Swarm Algorithms And Their Synergy

The novel approach introduced here is taking advantage of swarm intelligence to facilitate optimization process in TV regularization. An ARMA model used for the nonlinearly degraded image deconvolution is identified using ANN which is fast trained by two swarm algorithms hybridization: PSO and BFO. Both estimated image and blur function are identified through this representation. Some applications on radiological images are presented in simulation results. This optimized model will be implemented on reconfigurable hardware. Computational comparison based on image quality reached is made between different approaches.

Autorentext

Received his BSc in Elect.Eng.from INELEC(Algeria) and MSc degree in Signal Processing from Univ.of Blida (Algeria). PhD degree from SAAD DAHLAB univ. He worked as an indust.inst. Eng. in the Algerian power systems society for many years. He is currently with the Faculty of Sc.and Tech. in Djelfa univ. Algeria, as an Ass. Prof.