Adaptive region growing algorithm combined hue statistical characters
融合色调统计特征的自适应区域增长算法

This paper proposed a adaptive region growing algorithm HARG combined hue, hue frequency and its discrete curvature, to direct at the drawback of conventional region growing algorithm characterized by the seed selected and growing condition confirmed great effected and ineffective confirmed. It calculated seed using the sum of hue histogram added window fulfilled with threshold, and region growing by the threshold. Used the algorithm for single cell of microscope cell image, and the experiment results show that the HARG algorithm enhances segmentation precision and achieves segmentation of multiple regions by ensuring adaptively the seed and growing condition.