Threshold Selection Using Maximum Autocorrelation Principle
Qian Wang · 2011
The threshold selection method for image segmentation presented here is based on maximum autocorrelation principle. The optimal threshold value is determined by maximizing the autocorrelation measure. The predominance of our method is using both gray similarity and spatial continuity in object and background region of the image. The effectiveness of the thresholding results are demonstrated by its performance on some categories images.