Adaptation of ultrasound image texture characterization parameters

Statistical texture analysis has been commonly applied in the quantitative characterization of ultrasonic image. In this investigation, quantitative characterization of ultrasonic tissue was carried out using co-occurrence matrix parameters. The techniques involved were substantially improved as compared to the standard algorithm. Results were obtained from ultrasound images of 4 different tissue mimicking foam test objects under different gain settings. The performance of all the parameters was evaluated using the discriminant analysis. The standard algorithm of the spatial grey level concurrence matrix was also implemented for comparison purposes. The classification accuracy of the novel techniques can be as high as 96% in the training set and 94% in the testing set. The largest improvement over the standard algorithm is 13% in the training set and 14% in the testing set.