A Multilevel Image Thresholding Based on Hybrid Jaya Algorithm and Simulated Annealing
Bushra Abd, Azmi Tawfek Hussein Alrawi, Atheer Bassel · Journal of university of Anbar for pure science · 2023
Thresholding is a method for region-based image segmentation, which is important in image processing applications such as object recognition Multilevel. Thresholding is used to find multiple threshold values. Image segmentation plays a significant role in image analysis and pattern recognition. While threshold techniques traditionally are quite well for bi-level thresholding algorithms, multilevel thresholding for color images may have too much processing complexity. Swarm intelligence methods are frequently employed to minimize the complexity of constrained optimization problems applicable to multilevel thresholding and segmentation of color (RGB) images; In this paper, the hybrid Jaya algorithm with the SA algorithm was proposed to solve the problem of computational complexity in multilevel thresholding. This work uses Otsu method, Kapur entropy and Tsallis method as techniques to find optimal values of thresholds at different levels of color images as the target Tasks Experiments were performed on 5 standardized color images and 3 grayscale images as far as optimal threshold values are concerned, Statistical methods were used to measure the performance of the threshold methods and to select the better threshold, namely, PSNR (Peak Signal to Noise Ratio), MSE (Mean Square Error), SSIM (Structural Similarity Index), FSIM (Feature Similarity Index) and values of objective at many levels. The experimental results indicate that the presenter's Jaya and Simulated Annealing (JSA) method is better than other methods for segmenting color (RGB) images with multiple threshold levels. On the other hand, the Tsallis entropy of the cascade was found to be more robust and accurate in segmenting color images at multiple levels.