Image thresholding based on θ-division of 2-D histogram and maximum Shannon entropy

Abstract

In view of the obvious wrong segmentation in commonly used region division of 2-D histogram and the non- universality of oblique segmentation method for image thresholding proposed recently, in this paper a much more widely suitable thresholding method is proposed based on the θ-division of 2-D histogram and the maximum Shannon entropy criterion. Firstly, the θ-division method of 2-D histogram is given. The region is divided by four parallel oblique lines and a line, where the angle between its normal line and gray level axis is θ degrees. Image thresholding is performed according to pixel's weighted average value of gray level and neighbour average gray level. The oblique segmentation method can be regarded as a special case of the proposed method at θ=45°. Then the formulae and its fast recursive algorithm of the method are deduced. Finally the segmented results and the running time at different values of θ are listed, which show that the segmented images achieve more accurate borders at smaller values of θ and the anti-noise is better at larger values of θ. The value of θ can be selected according to the real image characteristics and the requirements of segmented results. Compared with the algorithm of conventional 2-D maximum Shannon entropy method, the proposed method not only achieves more accurate segmentation results and more robust anti-noise, but also reduces the running time and memory space significantly.