The current study is focused on an image segmentation algorithm for Uniformity Quality assessment in Diagnostic Ultrasounds.In particular a mathematical definition of the uniformity in ultrasound images is introducedand a split and merge algorithm performed on sparse matricesto measure uniformity is described. The algorithm is based on the Gray-Level Co-occurrence Matrices and the relativedescriptors,i.e. the Haralick features Entropy, Energy, Maximal CorrelationCoefficientand InformationMeasures of Correlation.Results on2 differentdatasetsof test images with different non-uniformities have been carried on. Several outcomesshow a goodsensitivity and agreement with the mean judgment by 7 human observers, i.e. differences are below 40% in most of the cases.On the basis of previousstudies, the latest developments and results are proposed and commented.
Schinaia, L., Scorza, A., Orsini, F., Sciuto, S.A. (2017). Ultrasound image uniformity assessment by means of sparse matrices: Algorithm implementation and first results. In 22nd IMEKO TC4 International Symposium and 20th International Workshop on ADC Modelling and Testing 2017: Supporting World Development Through Electrical and Electronic Measurements (pp.164-169). IMEKO-International Measurement Federation Secretariat.