Image Denoising of Medical Ultrasound Images using Curvelets
Conference: International Conference on Soft Computing Applications in Wireless Communication
Ultrasound imaging is most extensively used imaging method due to its clinical, inexpensive nature.\nUnfortunately, the quality of medical ultrasound images is usually restricted owing to a number of factors. The major\nproblem of ultrasound imaging is presence of speckle noise while acquisition. Speckle noise have tendency to decrease the\nimage contrast and distort details of image, thereby degrading the quality and consistency of ultrasound images. Due to these\nreasons, the difficulty in analysis process increases. Various methods have been used to suppress speckle in ultrasound\nimaging. Most popular methods are wavelet based transformation and curvelet transformation. Curvelet transform is a higher\ndimensional overview of the wavelet transform intended to illustrate images at various angles and scales. It is a transform\nwith multi-scale representation through various directions at each extent. This paper presents a novel algorithm for image\ndenoising in medical ultrasound imaging. The proposed algorithm provides an efficient way to use the threshold algorithms\nto enhance ultrasound images based on curvelet transform and also suppresses noise present in the ultrasound images. The\nquantitative and qualitative comparisons show that proposed algorithm outperforms other existing algorithms used for\nmedical ultrasound image denoising without blurring the edges and without causing over smoothing of detailed features of\nthe image. The metrics used for evaluation are Coefficient of Correlation (CoC), Edge Preservation Index (EPI), Structural\nSimilarity Index (SSIM) and Signal to Noise Ratio (SNR).
SCAWC - 2017