Enhancement of video demosaicing techniques using optimization algorithm

Abstract

SummaryDemosaicing and denoising are the most significant steps in the imaging process for both images and videos. The correlation noise corrupts the image quality, such that an accurate interpretation of the image is avoided. Thus, demosaicing is utilized to restructure the full‐color image from imperfect color samples. In this article, cat shuffled shepherd optimization‐based local polynomial approximation and intersection of confidence intervals (Cat SSO‐based LPA‐ICI) approach is developed for video demosaicing. The LPA‐ICI model is employed for frame demosaicing in which the second‐order polynomial interpolation filter coefficients are produced using the devised optimization algorithm, named cat SSO technique. Similarly, generative adversarial network classifier is also applied for frame demosaicing. In addition, initialization and flow estimation is performed for the demosaicing process. Besides, spatio temporal demosaicing model is utilized in order to get a noise‐free sequence. However, the developed cat SSO technique is newly devised by the incorporation of the shuffled shepherd optimization algorithm and cat swarm optimization technique. The created video demosaicing model outperformed with peak signal‐to‐noise ratio, root mean square error, and second derivative‐like measure of enhancement values of 56.85 dB, 0.46, and 60.62, respectively.