Multi-level image thresholding using GOA, WOA and MFO for image segmentation

Nowadays, in image segmentation algorithms, meta-heuristic algorithms are widely used todetermine multi-level thresholds. Many meta-heuristic algorithms use different methods asthe fitness function to determine multi-level thresholds. They may encounter prematureconvergence to determine the number of thresholds, and fail to obtain the correct answer,which in this case leads to an inaccurate image segmentation and may even lower the qualityof the image. In this paper, Moth-Flame Optimization (MFO), Whale OptimizationAlgorithm (WOA) and Grasshopper Optimization Algorithm (GOA) are utilized to determinemulti-level thresholds, which use a mathematical equation using the corresponding imagefeatures as a fitness function. According to the experiments, all three proposed algorithms forthe fitness function, it has a much better performance than the other algorithms and GOA isbetter than other algorithms and it has been able to increase the quality of image.

Keywords:

Image segmentation , multi-level thresholding , Moth-Flame Optimization (MFO) , Whale Optimization Algorithm (WOA) and Grasshopper Optimization Algorithm (GOA).