Adaptive F-SCA Method for Video Super Resolution

Abstract

This work presents an adaptive fractional-sine cosine algorithm (F-SCA) optimization technique for noise reduction and video super resolution. Adaptive F-SCA technique is the combination of Sine Cosine algorithm and the fractional calculus to make the F-SCA algorithm adaptive for better video enhancement. This technique is used for converting low resolution videos to super resolution (SR) videos. The proposed method consists of alignment filter, three-dimensional convolution system, batch normalized modules to perform image filtering and also to reduce the blurriness issue in low resolution frames. The performance of proposed method is measured in terms of Peak-SNR and SSIM value. Experimental analysis was conducted through CamVid Database. The performance of the Adaptive F-SCA is evaluated using PSNR, and SSIM. The suggested technique gives a maximum PSNR, SSIM values of 29.182 dB, 0.9366 respectively. Indicating that it is superior to other existing methods. The proposed hybrid technique outdoes the prevailing methods with PSNR value of 33.5026 dB, SDME value of 41.1859 dB and a maximum of 0.6222, SSIM value. The proposed method gives maximum PSNR improvement of ≈ 18% in comparison with existing literature. This technique is used for enhancing resolution of noisy videos to super resolution videos.