Gastroscopic image retrieval based on color-texture autocorrelation algorithm
Xiaoying Tai · Journal of Circuits and Systems · 2011
To extract the color and texture information of gastroscopic image effectively,and to incorporate spatial correlation feature,a method of gastroscopic image retrieval based on color-texture autocorrelation algorithm is presented in this paper.Firstly,the texture images extracted by texture spectrum and automatic thresholding are combined with color feature vector,then we calculate the color-texture autocorrelation vector at distance 1 and 3.Generalized Tversky Index model is used in retrieval for similarity measure.At last,we compare our method with other low-level feature extracting algorithms.Experimental results show that the method discussed in this paper is much more effective.