Entropy based thresholding of corss-dissolved ultrasound images

In this paper we present a new technique to improve the quality of entropy-based ultrasound image thresholding. The technique is based on the idea of cross-dissolving two images of the same object and thresholding the resulting image instead of thresholding any of the original images. We present the results of two experiments and how sharpening of the cross-dissolved image could improve results further. We performed our experiments on three entropy-based thresholding methods. The results obtained for each method are discussed.