Medical Image Retrieval Based on Bidimensional Empirical Mode Decomposition

Abstract

An approach of medical image decomposition and texture feature extraction based on the bidimensional empirical mode decomposition(BEMD), which can decompose the image into a set of functions denoted intrinsic mode functions (IMF) and a residue, was presented. Features extracted were the mean and standard deviation of the amplitude matrix, phase matrix and instantaneous frequency matrix of the IMFs and their Hilbert transformations. The extracted features were used for medical image retrieval. Moreover, according to the spatial relationship between local extrema points, a new boundary processing method based on clustering algorithm was proposed. In order to evaluate the proposed BEMD-based feature, we also presented a new multiscale fractal dimension feature. Preliminary comparison experimental results showed that the retrieval results of the BEMD-based feature were encouraged.