Image segmentation approach based on adaptive flower pollination algorithm and type II fuzzy entropy

Image segmentation depend on fuzzy entropy (FE) and intelligent optimization is among the most widely used and popular approaches. Segmentation is an important and pre-processing step in the analysis of an image. Multilevel thresholding is efficient for color images in different multimedia applications in day-to-day life. The method of assessing optimal threshold values using conventional schemes consumes more time. To alleviate the above-mentioned problem, meta-heuristic method has been used for optimization in this area over the last few years. This paper proposes a novel image thresholding technique depend on Adaptive Flower Pollination Algorithm (AFPA) and type II fuzzy entropy (TII-FE). The thresholding methodology is assessed against competitive algorithms concerning the quality, convergence and accuracy of segmented images. The quality is computed in relation of SSIM, PSNR and MSE parameters. The results indicate that AFPA for TII-FE is effective technique for image thresholding.

Automated summary

This paper proposes a novel image thresholding technique based on Adaptive Flower Pollination Algorithm and type II fuzzy entropy. Through the evaluation of quality, convergence and accuracy, the effectiveness of this technique in image segmentation is demonstrated.