Objective Methods for Quality Assessment of Digital Television Images for Broadcasting and Applied Television
Annotatsiya
This paper is an experiment study on how effective digital image resizing is using the different interpolation techniques. The emergence of digital broadcasting and used television triggered a significant increment in the need to broadcast high definition (HD) and ultra high definition (UHD) television programs, which has caused a considerable increase in the amount of video data. The trend brings about the issue of the possible deterioration of visual quality when bulk amounts of data are transferred in real time through the available communication channels.One of the main tasks of video processing is that images need to be scaled-both up (upscaling) and down (downscaling) in the transcoding process in order to ensure compatibility across different resolutions. Additionally, the transmission of video streams with a lower resolution followed by the expansion back to the original dimensions allows significant traffic savings. Therefore, the ability to resize an image adaptively, without changing its visual fidelity, is of significant practical and scientific importance.The research examined and experimentally tested commonly used interpolation techniques for image resizing, namely Bicubic, B-Spline, Mitchell and Lanczos. The experimental protocol involved reducing the sizes of test images successively by one half followed by restoration to the original size using the same interpolator. The quality and fidelity of the reconstructed images were assessed quantitatively through visual inspection of the images and quantitatively through mean square error (MSE). The article presents results which show the relative effectiveness of each of the interpolators investigated in the reduction and restoration operations.