Color Printing Image Quality Evaluation Method Based on Full Quaternion Matrix
| Autor: | Yunfei Zhong, Fu Lujing, Lu Chen, Ruojing Wang, Yaojian Hu |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Image quality
Computer science Color image business.industry Feature vector MathematicsofComputing_NUMERICALANALYSIS ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Color printing Singular value Singular value decomposition RGB color model Computer vision Artificial intelligence Quaternion business ComputingMethodologies_COMPUTERGRAPHICS |
| Zdroj: | Lecture Notes in Electrical Engineering ISBN: 9789811035296 |
| DOI: | 10.1007/978-981-10-3530-2_26 |
| Popis: | This paper proposed a color printing image quality evaluation method on the basis of the full quaternion matrix. First of all, according to the properties of the quaternion matrix and the characteristics of the color image, color image luminance component of local variance are used as a real part of quaternion and built color printing images of the quaternion representation model. Then to singular value decomposition of quaternion, calculate the quaternion matrix corresponding to the block of singular value feature vector, take the maximum value as the main parameter of measuring structural similarity images. Color printing image quality evaluation is realized by comparing a maximum of two images’ singular value feature vector. The experimental results show that the full quaternion method is used to express the structure of the color image information which highlight the human eye visual structure of sensitive information in the image quality assessment as well as realizes the RGB color image information in three-channel parallel processing, enhances and improves the efficiency and accuracy of color printing image quality assessment. At the same time the evaluation results conform to the human eye subjective visual perception. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|