Context-based Image Interpolation Method

The development of digital image processing technology and wide application are promoted by the rapid development of computer technology and the increasing awareness of the requirements of image quality, interpolation as an important method of image processing technology also has been widely used, interpolation method can be used to solve many problems in image processing field, such as image zoom, image de-interlacing, image denoising, image inpainting and so on. Traditional interpolation algorithms because of its low-pass filter characteristics, the high-frequency detail of image is lost, and these details are often very important. Hencely, the edge-based interpolation has been rapidly developed.The work in this paper are based on the theory of regression and autoregressive interpolation algorithm. Autoregressive interpolation problem of interpolation effect in smooth regions with high rough and high time complexity have been improved, while the application of regression model based on maintaining the edge in the image de-interlacing has been researched. The main research work are as follows:(1) Image interpolation based on the improved autoregressive algorithm is proposed. According to the characteristics of the correlation between the image pixels, the scope of the interpolation has been changed. Experimental results show that, the improved interpolation algorithm has improved the rough interpolation problem in smooth regions;(2) Image autoregressive interpolation algorithm based on edge detection is proposed. According to the next sampled edge detection image and with the interpolated restoration image and edge detection image mapped, the inerpolated pixel will be inserted into the edge pixel point and smooth point and the different points will be interpolated with different interpolation algorithms, then the complexity of the algorithm has been reduced. Experiments show that, the algorithm not only ensure the premise of the image visual quality, but also reduce the time complexity;(3) Image de-interlacing algorithm based on regression is proposed:the regression model based on maintaining the edge is applied for the image de-interlacing Experiments show that, the algorithm improves the PSNR value.