<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>First, this paper proposes the EPNN algorithm to recognize the high-resolution art of texture images. Then, a multi-scale multivariate image analysis method for color-textured surface classification is established. Regarding probabilistic neural networks, a differential evolutionary algorithm is proposed to optimize the smoothing parameters of basic probabilistic neural networks. Texture feature extraction involves extracting energy and statistical features of texture images using tree-structured wavelet packet decomposition and statistical-based methods. The image classification is performed by estimating similarity distances and classifying typical feature clusters. The results show that in terms of recognition ability, the proposed EPNN in this paper achieves an average improvement of 0.216 in CCP compared to the LBP method. In terms of classification ability, the proposed multi-scale multivariate image analysis color texture surface classification method in this paper generally achieves more than 10% higher classification ability than LGP in CUReT and Outex texture image libraries.</p>
</abstract>

First, this paper proposes the EPNN algorithm to recognize the high-resolution art of texture images. Then, a multi-scale multivariate image analysis method for color-textured surface classification is established. Regarding probabilistic neural networks, a differential evolutionary algorithm is proposed to optimize the smoothing parameters of basic probabilistic neural networks. Texture feature extraction involves extracting energy and statistical features of texture images using tree-structured wavelet packet decomposition and statistical-based methods. The image classification is performed by estimating similarity distances and classifying typical feature clusters. The results show that in terms of recognition ability, the proposed EPNN in this paper achieves an average improvement of 0.216 in CCP compared to the LBP method. In terms of classification ability, the proposed multi-scale multivariate image analysis color texture surface classification method in this paper generally achieves more than 10% higher classification ability than LGP in CUReT and Outex texture image libraries.

High-resolution art recognition of modern packaging design and images based on texture image classification

Applied Mathematics and Nonlinear Sciences

This work is licensed under the Creative Commons Attribution 4.0 International License.

First, this paper proposes the EPNN algorithm to recognize the high-resolution art of texture images. Then, a multi-scale multivariate image analysis method...

High-resolution art recognition of modern packaging design and images based on texture image classification

Online veröffentlicht: 01. Nov. 2023

Seitenbereich: -

Eingereicht: 06. Jan. 2023

Akzeptiert: 12. Mai 2023

DOI: https://doi.org/10.2478/amns.2023.2.00911

Schlüsselwörter
EPNN, Multi-scale multivariate images, Texture image classification, Modern packaging design, Art Recognition

© 2023 Jun Shi, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

High-resolution art recognition of modern packaging design and images based on texture image classification

Online veröffentlicht: 01. Nov. 2023

Seitenbereich: -

Eingereicht: 06. Jan. 2023

Akzeptiert: 12. Mai 2023

DOI: https://doi.org/10.2478/amns.2023.2.00911

SchlüsselwörterEPNN, Multi-scale multivariate images, Texture image classification, Modern packaging design, Art Recognition

© 2023 Jun Shi, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Schlüsselwörter
EPNN, Multi-scale multivariate images, Texture image classification, Modern packaging design, Art Recognition