Multi-level Image Thresholding with Global-Best Distance Artificial Bee Colony Algorithm

Segmentation is a very important step in image processing. One of the segmentation methods is multi-level thresholding. The success of multi-level thresholding depends on the best way to determine the appropriate threshold values. In this study, we tried to find the best threshold values by using Global-Best Distance ABC (G-BD ABC) algorithm which is one of the improved Artificial Bee Colon (ABC) algorithms. For the thresholding, Otsu's inter-class variance formula is used as a fitness function. Experiments were performed on 5 images by selecting various threshold values. The results are compared with the original ABC algorithm. When the results are examined, it is observed that the algorithm gives the same result when the threshold value is selected as two and the G-BD ABC algorithm gives the good result when the threshold value is increased.