A preliminary comparison of two different methods for objective uniformity evaluation in diagnostic ultrasound imaging

Although ultrasound image uniformity is a very important parameter for quality assurance in diagnostic ultrasounds, it is usually assessed by a qualitative and subjective judgement of technicians. In this work two novel method to obtain an objective measurement of the B-mode image uniformity over the whole field of view (or some of its part) are briefly described and compared: with the aim to quantify non-uniformities the first method is based on the image gray level histogram weighted by a sigmoid function (Sigma Weighted Histogram Method or SWHM) while the second one applies a segmentation of the Region of Interest, depending on some texture features from co-occurrence matrices processing (Texture Distribution Analysis Method or TDAM). Results from the two methods are preliminary compared and discussed on a set of 9 test images provided by means of two commercial ultrasound phantoms applied on different ultrasound scanners: the sensitivity to nonuniformities in SWHM is lower than in TDAM, on the other hand TDAM scores are more unstable and affected by higher uncertainties due to co-occurrence matrices calculations. Both methods require improvements and an in depth validation testing, nevertheless results are encouraging.