Image interpolation using interpolative classified vector quantization

Hong, Sungho; Park, Rae-Hong; Yang, Seungjoon; Kim, Junyong

doi:10.1016/j.imavis.2007.05.002

Cited by 19 publications

(9 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The edge-based category contains such as the locally adaptive threshold method [12], fast edge-oriented algorithm [13], edge-and-corner preserving [14], and soft edge smoothness [15]. The learning-based category includes such as the self-decomposed codebook [16], two-step interpolation based on training set [17], classified vector quantisation interpolation [18], and neural network-based method [19]. The above-mentioned methods were proposed for grey images.…”

Section: Introductionmentioning

confidence: 99%

“…In the meantime, there are also many methods proposed for color images such as the saliency-directed interpolation [20] and combining vector ordering interpolation [21]. In brief, in spatial domain, [9][10][11][12][13][14][15][16][17][18][19][20][21] cost more computation time than our approach on enlargement processing. In addition, most of methods [9][10][11][12][13][14][15][16][17][18][19] cannot deal with non-integer factor enlargement; in frequency domain, [5][6][7][8] do not consider global information in whole image, such as the traditional DCT method.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Fast large‐scale image enlargement method with a novel evaluation approach: benchmark function‐based peak signal‐to‐noise ratio

Guo¹,

Hsu²,

Kuo³

et al. 2015

IET Image Processing

View full text Add to dashboard Cite

An objective novel evaluation approach, implemented by the benchmark function-based peak signal-to-noise ratio, particularly suitable for evaluating the performance of a large-scale enlargement of a small size image is proposed in this study. Also, a fast large-scale image enlargement method via the improved discrete cosine transform (DCT) is proposed to improve the quality and speed of image zooming. The proposed image enlargement algorithm based on DCT saves computation time by multiplication of the DCT matrix. Compared with the traditional DCT approach, the improved approach overcomes the image shifting and blocky effects. In comparisons with other interpolation methods, DCT enlargement outperforms them in edge details because it considers the global frequency information of the whole image. With the DCT enlargement, it is easy to implement the arbitrary pixel-size-based zooming of an image by employing the different size of transform matrix. Illustrative examples show the effectiveness of the proposed approach.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Fast large‐scale image enlargement method with a novel evaluation approach: benchmark function‐based peak signal‐to‐noise ratio

Guo¹,

Hsu²,

Kuo³

et al. 2015

IET Image Processing

View full text Add to dashboard Cite

show abstract

“…A NURBS-based method is designed for interpolation of the simple components, and a distance map-based method for the complex components. Hong et al [25] presented image interpolation using interpolative classified vector quantization. To reduce the computational load, this algorithm reconstructs an interpolation image by adding to an initially-interpolated image high-frequency components predicted from training with a number of example image sets.…”

Section: Introductionmentioning

confidence: 99%

Research on Interpolation Methods in Medical Image Processing

2010

View full text Add to dashboard Cite

Image interpolation is widely used for the field of medical image processing. In this paper, interpolation methods are divided into three groups: filter interpolation, ordinary interpolation and general partial volume interpolation. Some commonly-used filter methods for image interpolation are pioneered, but the interpolation effects need to be further improved. When analyzing and discussing ordinary interpolation, many asymmetrical kernel interpolation methods are proposed. Compared with symmetrical kernel ones, the former are have some advantages. After analyzing the partial volume and generalized partial volume estimation interpolations, the new concept and constraint conditions of the general partial volume interpolation are defined, and several new partial volume interpolation functions are derived. By performing the experiments of image scaling, rotation and self-registration, the interpolation methods mentioned in this paper are compared in the entropy, peak signal-to-noise ratio, cross entropy, normalized cross-correlation coefficient and running time. Among the filter interpolation methods, the median and B-spline filter interpolations have a relatively better interpolating performance. Among the ordinary interpolation methods, on the whole, the symmetrical cubic kernel interpolations demonstrate a strong advantage, especially the symmetrical cubic B-spline interpolation. However, we have to mention that they are very time-consuming and have lower time efficiency. As for the general partial volume interpolation methods, from the total error of image self-registration, the symmetrical interpolations provide certain superiority; but considering the processing efficiency, the asymmetrical interpolations are better.

show abstract

“…While effective, [1] and other related algorithms [2] is primarily designed as a regression technique rather than a signal processing approach, where superresolution and image interpolation algorithms are traditionally developed.…”

Section: Introductionmentioning

confidence: 99%

Empirical Type-i filter design for image interpolation

Nguyen

2010

2010 IEEE International Conference on Acoustics, Speech and Signal Processing

View full text Add to dashboard Cite

Empirical filter designs generalize relationships inferred from training data to effect realistic solutions that conform well to the human visual system. Complex algorithms involving multiple linear regressions produce optimal results, but a single zero-phase filter yields comparable image quality at a fraction of the computational load. We propose an algorithm that builds a single symmetrical linear filter based purely on collected training data. Such a filtering technique balances the tradeoff between performance and complexity. Previous implementations of zero-phase interpolation filters as well as other learning-based interpolating algorithms are analyzed and examined. The proposed algorithm utilizes a Type-I symmetrical filter, an improvement and alternative over previous work on Type-II empirically-based interpolating filters. Given image training patches, the work discusses the enforcement of our filter properties while simultaneously drawing information from the training set. Additionally, we describe the implementation of the designed filter, its application, and related considerations. Finally, advantages of the proposed algorithm are analyzed.

show abstract

Image interpolation using interpolative classified vector quantization

Cited by 19 publications

References 25 publications

Fast large‐scale image enlargement method with a novel evaluation approach: benchmark function‐based peak signal‐to‐noise ratio

Fast large‐scale image enlargement method with a novel evaluation approach: benchmark function‐based peak signal‐to‐noise ratio

Research on Interpolation Methods in Medical Image Processing

Empirical Type-i filter design for image interpolation

Contact Info

Product

Resources

About