Modified BPS algorithm based on shearlet transform for noisy images

Most of the medical and nuclear images contain visual noise. Due to the presence of noise, the images can't be examined properly. So, there is a necessity to avoid the noise in order to provide better quality images along with the better compression efficiency for efficient storage and transmission. Block Based Pass Parallel SPIHT (BPS) is widely used for compression because of its high processing speed. It is possible by decomposing the wavelet transformed image into 4×4 bit-blocks and the decomposed blocks are encoded simultaneously in all the bit planes hence the speed is very high. However, the major drawback is the slight degradation in the PSNR value and visual quality. To overcome this drawback, a modified BPS algorithm is proposed which replaces wavelet with Shearlet because shearlet provides multi directional information and it also used to detect the geometrical features like edges. LLSURE technique is appliedbefore transformation to remove the noise induced in the image. It is preferred because of its high edge preserving capability. Experimental results demonstrate the effectiveness of the LLSURE filter and the shearlet transform in the BPS algorithm. It shows that the PSNRvalue is very effective for images corrupted with Gaussian noise when compared to other noisy images.