Study of image magnification based on Newton interpolation
Jiang Tao, Wenhu Lu, Dong Shouyin, Han Mingqin
Abstract
Jiang Tao, Wenhu Lu, Dong Shouyin, Han Mingqin
Abstract
The common methods of image magnification mainly included nearest neighbor, bilinear interpolation. However, it would appear obvious sawtooth shape in the high-frequency part of the magnified image by nearest neighbor method, and become fuzzy by bilinear interpolation method, all of above could affect the image quality. Compared with the above two methods, this paper found a new method using Newton interpolation for image magnification, and used the Matlab to implement this algorithm. The experiments showed that this algorithm could save more high frequency information of the image based on the characteristics of pixel and reduce the fuzzy degrees of the image. Meanwhile, this paper gave an objective evaluation in the image's entropy of information and average gradient that this method better than conventional magnification method, so it could be widely used in the field of engineering practice.
OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
The common methods of image magnification mainly included nearest neighbor, bilinear interpolation. However, it would appear obvious sawtooth shape in the high-frequency part of the magnified image by nearest neighbor method, and become fuzzy by bilinear interpolation method, all of above could affect the image quality. Compared with the above two methods, this paper found a new method using Newton interpolation for image magnification, and used the Matlab to implement this algorithm. The experiments showed that this algorithm could save more high frequency information of the image based on the characteristics of pixel and reduce the fuzzy degrees of the image. Meanwhile, this paper gave an objective evaluation in the image's entropy of information and average gradient that this method better than conventional magnification method, so it could be widely used in the field of engineering practice.
Key concepts: Bilinear interpolation, Image scaling, Magnification, Stairstep interpolation, Interpolation (computer graphics), Bicubic interpolation, Nearest-neighbor interpolation, Demosaicing