Multi-level image segmentation based on an improved differential evolution with adaptive parameter controlling strategy

Multi-level threshold segmentation techniques are one of the most important parts in image processing. They are simple, robust, and accurate. However, some of them have long computation time and it grows exponentially with the number of thresholds increase. This paper proposed an improved differential evolution with novel mutation strategy and adaptive parameter controlling method (MApcDE) so as to avoid time-consuming and overcome the relation between computation time and dimensions. OTSU method, which maximizes the variance between foreground and background in an image, is a popular threshold image segmentation technique, and is used in this paper to test the performance of the proposed method. Experimental results show that our proposed MApcDE algorithm can get more effective and preferable results when compared with some other population-based threshold methods. The computation time is shorten at the same time.