Comparison of the segmentation of ultrasonic image utilizing different fractal parameters

摘要

Abstract: Statistical texture analysis has been commonly applied in the quantitative characterization of ultrasonic image. Recently, another approach begins to emerge, which uses fractal dimension. In this investigation, fractal dimension is estimated from histologically confirmed ultrasonic images using three different approaches: intensity- based, spectrum-based and reticular cell counting. The parameters are then used in image segmentation. The fractal model has the advantages that the parameter generated is stable over transformations of scale and linear transforms of intensity. From the results obtained, the performance of fractal dimension obtained by the reticular cell counting method is better than the other two approaches and comparable to the spatial grey level co-occurrence matrices statistic.!11