Augmented echocardiography based on a hybrid computational strategy

Ultrasound (US) is the basic physical principle that, in the medical imaging context, allows the generation of ultrasound images of various structures of the human body. Usually, two-dimensional transthoracic echocardiography (ET2) is a technique available in most health care centers and can generate, non-invasively and inexpensively, images of almost all regions of the heart. However, the information contained in these graphic representations is affected by imperfections that affect its quality, among which the speckled ultrasonic noise stands out. Such imperfections, on the one hand, constitute a real challenge for those who must analyze the ET2 images and, on the other, open the possibility of searching for alternatives aimed at increasing the quality of the information obtained in ultrasound by resorting to variants of imaging US modalities as well as the use of other more complex imaging techniques or digital processing images (DPI) techniques. This article considers the application of a hybrid DPI based on similarity and a white hat-top transform that adequately conditions ET2 images, allowing a better visualization of the anatomical structures present in them. This fact may contribute, preliminarily, in two directions: a) Reduce the need to consider US variants such as, for example, contrast US, or use other imaging modalities that, although they do improve the appearance of the images, are more expensive options that affect the economic resources of patients and health centers. b) Provide cardiologists with better-looking images that allow them to obtain useful descriptors in the diagnosis of various heart diseases.