Speckle noise reduction in echocardiographic images of aortic valve and cardiac chambers

Echocardiography is very commonly utilized in the diagnosis of cardiac abnormalities but it being an ultrasound based imaging modality, comes with inherent curse of speckle noise. It is proposed to test and analyze the effect of logarithmic based posterior sampling Bayesian estimation (PSBE) for denoising of transthoracic echocardiographic (TTE) images of aortic valve. PSBE based experimentations reveal that, the performance of technique degrades considerably, when TTE images are simulated with speckle noise and then denoised. To address this issue, a spatially adaptive Wiener filter is being embedded into logarithmic PSBE method known as hybrid PSBE. Proposed experimentations are performed using TTE images acquired in 2 windows with 5 different views, each providing different piece of information necessary for the accurate diagnosis of underlying abnormalities. The usefulness of denoising is evaluated based upon edge preservation, image quality index, and similarity index in addition to PSNR, SNR and MSE comparisons. The exhaustive visual and parametric analysis shows that the proposed hybrid techniques are more robust and effective in speckle noise reduction. The denoised images are enhanced using Butterworth low pass filter, which will assist the doctors in delineating true cardiac chamber and valvular boundaries accurately with more ease and lesser intrapersonal variations.