Minimum Error Thresholding Based on Two Dimensional Histogram

Thresholding is an important technique for image segmentation and is widely applied to the field of digital image processing. The purpose of this paper is to extend the minimum error thresholding to the 2-dimensional histogram to come up a new algorithm, namely, minimum error thresholding based on the 2D histogram. This new method can derive the 2D optimal threshold vector by presuming that the object and background gray level values are normally distributed in the 2D histogram. This method utilizes not only the gray level information of each pixel but its spatial correlation information of its neighborhood in an image. The effectiveness of the algorithm is demonstrated by simulation results.