A fused contextual color image thresholding using cuttlefish algorithm

Abstract

In this paper, we have proposed a fusion-based context-sensitive Masi energy curve model for multi-level thresholding exploiting cuttlefish algorithm (CFA). The proposed algorithm is simple and very efficient for the task of color image segmentation. Although Masi entropy exploits the additive/non-extensive information with the aid of a concordant entropic parameter, the performance is observed to be poor in the case of color image segmentation. Improved results can be obtained by using the concept of energy curve with Masi entropy at the cost of increased computational cost while selecting the suitable thresholds. To overcome the aforementioned drawbacks as well as to increase the quality of the segmented image, a simple multi-level thresholding method is proposed in this paper. The proposed color image segmentation scheme exploits the concept of local contrast fusion along with CFA to resolve the aforementioned issues. In order to prove the effectiveness of the proposed scheme, experimental evaluations on standard daily-life color images have been reported in this paper. The experimental outputs demonstrate that fusion-based multi-level thresholding is better than the existing dominant segmentation methods.