Adaptive brightness transfer functions in echocardiography.

ABSTRACT

Despite the clear advantages of echocardiography as a diagnostic tool, its images tend to be noisy and unclear. This paper presents an innovative algorithm, called ABTF (adaptive brightness transfer function), designed to optimally adjust the gray-levels used in echocardiography. The algorithm is aimed at aiding in visual tissue classification and texture-based visual tissue tracking in echocardiographic images. The ABTF method is based on fitting the cine-loop's gray-level histogram to a sum of three Gaussian functions, each of which relates to a different region within the image, the left ventricular cavity, the relatively dark regions within the cardiac muscle and the bright regions within the cardiac muscle. The procedure's feasibility has been supported by a test-set, including 23 echocardiographic cine-loops from 10 different patients. The resulting image quality appears to be superior to that of the original images, tending to show better contrast and a higher dynamic range of gray-levels within the cardiac muscle. According to two expert cardiologists, who have blindly ranked the image quality of each cine-loop on a scale from 1 to 10, where 10 corresponds to the highest possible image quality, the mean score of the original cine-loops is 7.1 +/- 1.1, while the mean score of the cine-loops to which ABTF has been applied is 8.0 +/- 1.2.