Automatic threshold of image segmentation using 2-D entropy
Yunfei Zhang, Zhang Ye · Harbin Gongcheng Daxue Xuebao/Journal of Harbin Engineering University · 2006
Thresholding is an important form of image segmentation and is used in many applications that involve image processing and object recognition.Thus,how to acquire a threshold of image segmentation is key.To realize automatic segmentation and build a 2-D histogram using 2-D entropy,traditional methods research the central gray value and mean value of pixels in the same neighbor.However,the horizontal or vertical change of the grayscale value cannot represent the total characteristics among neighbor pixels very well.A new method is proposed to build a 2-D histogram by utilizing the center gray value and mean value of four other pixel gray values except four-neighbor pixels in the same 3×3 neighborhood.Experimental results on a classical image show the effectiveness of the new method.