Enhancing Image Segmentation: Multilevel Thresholding Using Artificial Bee Colony Algorithm

Thresholding is a sort of segmentation in which the pixels are divided into distinct groups based on their intensity level in accordance with one or more threshold values. Thresholding is a sort of segmentation in which the pixels are divided into distinct groups based on their intensity level in accordance with one or more threshold values. Thresholding is a popular image segmentation technique for converting gray-level images to binary images. This paper will demonstrate a Multilevel Thresholding (MTH) for image segmentation based on Artificial Bee Colony (ABC) algorithm. This paper shows the designing process for an image segmentation module using multilevel thresholding. The multilevel thresholding algorithm divides pixels into discrete zones, that segment the objects in the image, and it is the greatest solution for segmenting real-world pictures. The threshold point is established using the grey level histogram. It demonstrates how every image has a unique set of ideal threshold values. From there, we can get the optimum threshold value for multilevel thresholding based on the Artificial Bee Colony (ABC) algorithm. To get the optimum threshold value, Otsu’s method has been used. After applying Otsu’s Method, the algorithm is then analyzed by comparing the performance of multilevel thresholding using well-known benchmarks, Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM). The result shows that the ABC algorithm performs almost the same as the other three methods, which are Harmony Search (HS), Differential Evolution (DE) and Particle Swarm Optimization (PSO). This can be seen from the results of PSNR and SSIM which show small margins of difference between the four algorithms. It can be concluded that the ABC algorithm is proven to perform well in image segmentation. From the collected data, there are significant changes in the PSNR and SSIM values due to the various multilevel thresholding techniques. As the level rises, so does the quality of the segmented image.