Image segmentation using context sensitive multilevel thresholding method and hybrid meta-heuristic algorithms

Many traditional techniques histogram-based threshold for effective two-level thresholding and not be able to consider the contextual information space of the image to selectthe optimal threshold. In this paper a new thresholding method with an energy function to generate energy curve of an image bytaking account of contextual information space is provided, whose behavior is very similar to histogram. To incorporate contextual information spatial of the image for the selectionprocess ahead, the energy curve uses the function of the input of energy instead of histogram. In addition, to reduce the problemof multi-level thresholding, we use characteristics of the hybrid genetic algorithms (GA) and particle swarm optimizationoperation (PSO). This algorithm in a number of different types of images was evaluated using validity measure. The result of theproposed method using energy curve is compared with the image histogram based and context sensitive technique with genetic algorithm. These comparisons emphasized the effectiveness of theproposed method with Context sensitive technique with genetic algorithm. These comparisons emphasized the effectiveness of the proposed method.