Optimized Bayesian NL-means Blockwise approach for Ultra Sound Images
Mohammed Saleem Ahmed, Raju Korandla, SNVASRK Prasad, Anandbabu Gopatoti · International Journal of Engineering & Technology · 2018
Ultrasound imaging is a portable medical diagnostic tool provides real time clear images of soft tissues compared to x-ray imaging in which tissues are not visible up to the mark. Other diagnostic tools like Magnetic Resonance Imaging and Computer Tomography can well visualize the tissues but are cost effective. Due to the presence of speckle noise Ultrasound images gets degraded leads to low quality imaging. This noise down the spatial, contrast resolutions and Peak Signal to Noise Ratio (PSNR) in US images. Consequently, filtering techniques for speckle noise reduction are of unique interest for clinical ultrasound imaging. To reduce the speckle noise in US images, Blockwise scheme is adapted to the Non Local (NL) means filter based on Bayesian formula here in this paper. We proposed this Blockwise scheme to a NL-means filter to develop Optimized Bayesian NL-Means (OBNLM) filter which is suitable for removing speckle noise in US images. The experimental qualitative and quantitative result proves the proposed method is better than the NLM filter method. Also OBNLM filter keeps shape of original images with edges.