چكيده به لاتين
Abstract:
Various modifications have been suggested in the past to extend Shannon entropy to continuous random variables. This thesis investigates these modifications in the first chapter, and suggests a new entropy measure with the name of average entropy (AE) in the second chapter. AE is more general than Shannon entropy in the sense that its definition encompasses both continuous as well as discrete domains. According to the information theory, the concept of entropy is introduced as a branch of image processing by estimating the amount of information of an image that can provide a good level of information to describe the image and as a decomposition scale, the image information is divided into two or more of the two connected areas. In this case, it can be calculated from the distribution of the gray levels of the image and a good segmentation is obtained for the image. For this reason, for Effect of entropy on image segmentation, we need to have a proper understanding of the concept of image processing and its segmentation, so the third chapter is devoted to this topic. In the fourth and fifth chapters, this new measure and other four entropies are tested for their effectiveness in two images mamogram segmentation. The results show that in both images, the average entropy has better performance and greater accuracy.
Keywords: Information Theory, Shannon entropy, extend Shannon entropy, Average entropy, Image segmentation.
