Search Based Optimization Approach for Video Super Resolution

Abstract

Video Super Resolution (SR) is used to enhance the resolution of low resolution videos in minimum cost as compare to hardware models and it should handle noisy and blurry videos. This paper represents efficient and robust video SR model which enhance the resolution of different videos. Input video is taken from readily available standard UCSD datasets; these videos are in RGB format which are initially converted to HSV format. To achieve high resolution the V-channel is used for enhancement.  Full search motion estimation is used find out matching macro blocks in a video frame. Cubic spline interpolation technique is used to find value of unknown pixel. Bilateral total variation is used for de-noising and de-blurring. Resolution factor is optimized using different optimization techniques like particle swarm optimization (PSO), Grey Wolf optimization (GWO), Whale optimization algorithm (WOA) and Lion algorithm (LA). Performance of proposed video SR is compared with existing methods and statistical analysis is done with PSNR, SSIM, SDME, ESSIM and BRISQUE parameters.