Study And Analysis Of Speckle Noise Reduction Techniques For Ultrasound Images

Abstract

newline Ultrasound imaging is a boon technology in medical imaging with some limitations as the poor newlinequality of imaging due to speckle noise. Due to this behavior, Doctor/Radiologist feel newlineuncomfortable in diagnosis of a patient because of sometimes clinical details looks like speckle newlineimportant in ultrasound imaging. For this purpose, we have studied various state-of-the-art newlinetechniques related to this issue. We have done quantitative analysis of four filters: Median Filter, newlineLEE Filter, FROST Filter, and SRAD filter and found that SRAD filter is more efficient than newlineothers. After that, we have proposed a hybrid filtering technique with PSO to obtain a better newlinedespeckling result with minimized over filtering issues than other filtering techniques as newlineAdaptive Wavelet Thresholding and Joint Bilateral (AWT+JB) Filter, Adaptive Fuzzy Switching newlineWeighted Mean (AFSWM) Filter, Median Patch-Based Locally Optimal Wiener (MPBLOW) newlineFilter in terms of evaluation parameters PSNR, SNR, SSIM, and EPI. Filtering performance of newlinediscussed filters have shown in graphical form for better qualitative observation. Then, we have newlineproposed one more despeckling technique: robust denoising technique for ultrasound images by newlinesplicing of LRF and PCA. This technique can reduce speckle noise effectively and preserve the newlinediagnostic details of the ultrasound image up to a desired level. It performs better than Fast newlineFourier Bessel steerable PCA (FFBsPCA) Denoised filter, Two Phase Matrix Decomposition newlineDenoised filter, and New Adaptive Denoised filter for given ultrasound test images which have newlineresulted in terms of output ultrasound images, PSNR, and MSE graphs. These performance index newlinevalues have shown that proposed robust spliced denoising technique have substantial newlineachievement in enhancing the overall quality of ultrasound images.