摘要：Second-order derivatives are used to obtain more information from the pixels in a magnetic resonance image. This information is the difference between the values of pixels, which is a factor in the measurement of the level of contrast in the image. This information is presented in the form of a histogram of second-order values, which expresses differences in pixel values caused by contrast between details in the image. The proposed method of contrast measurement utilizes a Laplace curve together with the histogram, both of which are generated with image data. The curve and histogram are used to determine a measurement quantity that expresses the level of contrast and the presence of details in the image. The measurement quantity, which is a score, describes the change of contrast in the image as contrast enhancement is applied. This score is compared with other established contrast scores that utilize data from an individual image, without comparison to another image. These other scores are average local contrast, root mean square contrast and Michelson contrast. The goal of the comparison is to determine the performance of the proposed score at detecting the occurrence of excessive gain as contrast enhancement is applied on an MRI image. Sample MRI images, including images from established databases, are used for the testing and comparison. The proposed score performs better than average local contrast and root mean square contrast and provides results that do not diverge as gain is increased.
关键词：second-order derivatives; histograms; Laplace distribution curves; contrast measurement
