Echocardiography image enhancement using adaptive fractional order derivatives

摘要

Medical ultrasound images are low contrast in nature. Information regarding tissues and other important structure is required by a physician to assess patient's health. Therefore image enhancement is a critical pre-processing task. In this paper a methodology based on emerging topic of fractional calculus is proposed. Proposed method is simple yet effective. In the proposed algorithm, input image is first divided into smooth, texture and edge regions using gradient magnitude of each pixel. Then appropriate order of fractional differential mask is selected to enhance each pixel. Proposed method is compared with state-of-the-art histogram equalization method and fixed-order fractional differential methods. Results are verified quantitatively and qualitatively. For quantitative analysis average gradient and entropy are used. Simulation results verify the effectiveness of proposed method.