PSO based Optimization of Tuning Parameter for Quadrilateral Maximum Likelihood Estimation Despeckling

Image denoising has become a very essential exercise all through the diagnosis especially in case of medical image processing involving ultrasound. Speckle pattern is one of the most complicated multiplicative noise which hardly degrades visual quality of clinical ultrasound images. Real time ultrasound images posses inherited speckle pattern which reduces its resolution and contrast there by degrading the diagnostic accuracy of the ultrasound image. The presence of speckle noise in fetal ultrasound images make the conditions worse to carry out prenatal diagnosis of congenital heart disease. This is due to the impact of edge and local fine details that are not very clear for diagnosis. There exists an ever growing research demand for contriving a robust speckle reduction filter to enhance the quality of the speckle affected image and to preserve the essential features. This paper describes about the proposed despeckling filter contrived with the concept of Particle Swarm Optimization of weight parameters to improve the algorithm of Quadrilateral Rayleigh Maximum Likelihood Estimator to suppress the speckle noise in clinical ultrasound images. The proposed filter establishes the Rayleigh Maximum Likelihood Estimator and particle swarm optimization technique hence called as PSO based QRML filter. The experimental result shown in this paper proves the efficacy of the proposed filter in comparison with several existing despeckling filters in terms several performance indices and image profile. Experimental results shows that the proposed filter removes the speckle noise effectively and thus outshine the conventional filters.