An Improved Cuckoo Search Algorithm for Multithreshold Image Segmentation

Image segmentation is an important part of image processing. For the disadvantages under multiple thresholds such as long time and poor quality, improved cuckoo search (ICS) proposed for multithreshold strategy. Firstly, model based on maximum entropy threshold described, secondly, algorithm by using chaotic initialization population to improve diversity solutions, optimizing step size factor possibility obtaining optimal solution, probability reduce complexity algorithm; finally, function in used individual fitness solving. The simulation experiments show that has a good effect four different thresholding conditions.