2014 | OriginalPaper | Buchkapitel
An Efficient Method for Speckle Reduction in Ultrasound Liver Images for e-Health Applications
verfasst von : Suganya Ramamoorthy, Rajaram Siva Subramanian, Deebika Gandhi
Erschienen in: Distributed Computing and Internet Technology
Verlag: Springer International Publishing
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
There are seemingly an endless number of possible applications of information technology to health service management. The area of e-health is very broad, covers topics such as telemedicine, healthcare score cards, audits and information systems etc. In this paper, the focus is on identifying liver diseases at non invasive & low cost ultrasound modality. Our ultimate goal is to improve gray scale ultrasound images by removing speckles present in it. A new speckle noise reduction and image enhancement method, i.e. Laplacian pyramid nonlinear diffusion with Gaussian filter (Modified LPND), is proposed for medical ultrasound imaging. In the proposed Modified LPND, a coupled modified diffusivity function and gradient threshold is applied in laplacian pyramid domain of an image, to remove speckle and retains subtle features simultaneously. The performance of the modified LPND method is evaluated by both CNR and PSNR on a real ultrasound image dataset. In this work, we obtained an accuracy 95% and 12.34 in contrast-to-noise ratio and 13.32 in PSNR for liver cyst compared to the speckle reducing anisotropic diffusion (SRAD), nonlinear diffusion (ND) and LPND respectively. Also the proposed modified LPND showed clearer boundaries on both focal and diffuse ultrasound liver dataset. These preliminary results indicate that the proposed modified LPND can effectively reduce speckle noise while enhancing image edges for retaining subtle features like cyst and lesions.