Development of a new speckle reduction algorithm for medical ultrasound images

Among the various medical imaging for clinical diagnostic ultrasound plays a very vital role in clinical diagnosis. Ultrasound images suffer from an intrinsic artifact called speckle. Speckle degrades spatial and contrast resolution and obscures the underlying anatomy. This thesis aims to present an algorithm for suppressing speckle noise from medical ultrasound images in the wavelet domain. Wavelet thresholding and Bilateral filtering methods are applied to remove speckle noise from the ultrasound image. In wavelet thresholding algorithm, a suitable wavelet thresholding technique is decided for an effective speckle reduction in transform domain and selected threshold value is applied to the all wavelet coefficients. In this study, we also used Bilateral filter for speckle reduction. To enhance the performance of the the Bilateral Filter, and to preserve the edges, the Bilateral Filter is applied to the all approximation coefficients (LL) and detail coefficients (LH, HL, HH). The performance of the Wavelet thresholding and Wavelet Bilateral filter are compared with the state of the art methods. Experimental results show that Wavelet Bilateral filter exhibits better performance than the other methods. Standard measures like Mean Squared Error (MSE), Structural Similarity Index (SSIM), Peak Signal to Noise Ratio (PSNR) and running time are utilized to quantify the performance of the presented method.