Abstract | ||
---|---|---|
Originally inspired by the imaging needs of UND's Open Prototype for Educational Nanosats (OPEN) satellite program. The goal of this research is to direct the choice of an image interpolation/zoom algorithm.The image is scanned and overlapping 3x3 blocks of pixels are analyzed looking for “L” patterns. Such a pattern indicates that the value of the center pixel be changed transforming the “L” pattern into a triangle pattern. We compare this approach against different types of single-frame image interpolation algorithms, such as zero-order-hold (ZOH), bilinear, bicubic, directional cubic convolution interpolation (DCCI) approach, and a neural network approach. We use the peak signal-to-noise ratio (PSNR), mean squared error (MSE), Structural Similarity Index (SSIM), and Feature Similarity Index (FSIM) as the primary means of comparison. Tests cases include gray-scale and RGB images. In all test cases the proposed method resulted in a lower MSE, higher PSNR, and higher SSIM than the other zoom methods applied. The FSIM returned a higher value in every case for the ZOH method, but this was only when FSIM was calculated to the fourth decimal. Overall, this method results in a more accurate image after zooming than the other methods. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1109/EIT.2019.8834287 | 2019 IEEE International Conference on Electro Information Technology (EIT) |
Keywords | Field | DocType |
FSIM,image zooming,grayscale,color images,triangle pattern,single-frame image interpolation algorithms,neural network approach,peak signal-to-noise ratio,RGB images,structural similarity index,feature similarity index,PSNR,SSIM,Open Prototype for Educational Nanosats,OPEN satellite program,mean squared error | Pattern recognition,Computer science,Interpolation,Bicubic interpolation,Zoom,Electronic engineering,Artificial intelligence,RGB color model,Pixel,Image scaling,Grayscale,Bilinear interpolation | Conference |
ISSN | ISBN | Citations |
2154-0357 | 978-1-7281-0928-2 | 0 |
PageRank | References | Authors |
0.34 | 5 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ronald Marsh | 1 | 31 | 7.89 |
Md Nurul Amin | 2 | 0 | 0.34 |