SK-SVR: Sigmoid kernel support vector regression based in-scale single image super-resolution

Abstract

Learning based image super-resolution (SR) has been a striking area of research for generating high-resolution (HR) images from low-resolution (LR) images. A new in-scale single image super-resolution approach is proposed in this paper. The proposed approach effectively applies support vector regression (SVR) for learning and generates high resolution image. Contrasting to many learning based SR algorithms; the proposed approach does not require any training dataset in advance. In addition, sigmoid kernel SVR is used for generating error models and Bayesian decision theory is applied to select the model with the least errors. The performance of the proposed approach is evaluated in terms of peak signal-to-noise ratio (PSNR) and compared with state of the art learning based single image SR algorithms. The experimental results show that the proposed approach outperforms the other SR algorithms.