Variational Mode Decomposition based Image Segmentation using Sine Cosine Algorithm

Multilevel thresholding utilizing the histogram-based method is the well-known and preferred technique of image segmentation. This method generally suffers from irregularities and sharp details in the histogram, which leads to stagnation. Along with this, the computational time of this method grows exponentially as the number of thresholds increments. In this study, a freshly developed Sine Cosine algorithm is coupled with variational mode decomposition (VSCA) to overcome these problems. VSCA employed Masi entropy as a cost function for image segmentation. The outcome of the proposed algorithm examined with Non-VMD based method using quantitative parameter such as structural similarity index (SSIM), feature similarity index (FSIM), quality index based on local variance (QILV), computational time, and mean square error(MSE). The outcomes confirm that the proposed algorithm performs more reliable results than Non-VMD based techniques.