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

<p>In this paper, the color image is converted to a grayscale image in the image recognition preprocessing stage to accelerate the image recognition processing, and then the image contrast is enhanced by grayscale stretching to compute the grayscale layer covariance matrix and image texture features. Multi-step Markov clustering method is proposed to optimize the GCN, and the instance normalization layer and batch normalization layer are added to strengthen the source domain representation ability of the GCN to form a cross-domain image recognition algorithm based on a pairwise generalization network. Elaborate the visual design path of image information by artificial intelligence image recognition and image processing technology, introduce image recognition technology into the field of visual design, establish a visual design partition model, and completely extract the local feature information of computer image graphics. Use the dataset to evaluate the performance of pairwise generalization networks, and conduct simulation experiments to analyze the visual expression effects of visual design. The PGN-RM method, with the addition of maximum mean distance, instance normalization, and batch normalization, is able to achieve a performance average of 91.843. The peak signal-to-noise ratio of the actual effect image of the visual design of the product packaging is maintained in the range of [95.0312, 97.0032], which is an excellent visual design effect. Visual design that utilizes artificial intelligence graphic recognition technology can express design ideas more deeply and enhance the visual design’s attractiveness.</p>
</abstract>

In this paper, the color image is converted to a grayscale image in the image recognition preprocessing stage to accelerate the image recognition processing, and then the image contrast is enhanced by grayscale stretching to compute the grayscale layer covariance matrix and image texture features. Multi-step Markov clustering method is proposed to optimize the GCN, and the instance normalization layer and batch normalization layer are added to strengthen the source domain representation ability of the GCN to form a cross-domain image recognition algorithm based on a pairwise generalization network. Elaborate the visual design path of image information by artificial intelligence image recognition and image processing technology, introduce image recognition technology into the field of visual design, establish a visual design partition model, and completely extract the local feature information of computer image graphics. Use the dataset to evaluate the performance of pairwise generalization networks, and conduct simulation experiments to analyze the visual expression effects of visual design. The PGN-RM method, with the addition of maximum mean distance, instance normalization, and batch normalization, is able to achieve a performance average of 91.843. The peak signal-to-noise ratio of the actual effect image of the visual design of the product packaging is maintained in the range of [95.0312, 97.0032], which is an excellent visual design effect. Visual design that utilizes artificial intelligence graphic recognition technology can express design ideas more deeply and enhance the visual design’s attractiveness.

Research on Image Recognition and Visual Design Based on Artificial Intelligence

Applied Mathematics and Nonlinear Sciences

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

{"article-title":"Research on Image Recognition and Visual Design Based on Artificial Intelligence"}

In this paper, the color image is converted to a grayscale image in the image recognition preprocessing stage to accelerate the image recognition processing,...

Research on Image Recognition and Visual Design Based on Artificial Intelligence

Online veröffentlicht: 30. Mai 2024

Seitenbereich: -

Eingereicht: 12. Feb. 2024

Akzeptiert: 24. Apr. 2024

DOI: https://doi.org/10.2478/amns-2024-1236

Schlüsselwörter
Grayscale stretching, Multi-step Markov clustering, Cross-domain image recognition, PGN-RM, Visual design

© 2024 Yi Zhang, published by Sciendo

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

Research on Image Recognition and Visual Design Based on Artificial Intelligence

Online veröffentlicht: 30. Mai 2024

Seitenbereich: -

Eingereicht: 12. Feb. 2024

Akzeptiert: 24. Apr. 2024

DOI: https://doi.org/10.2478/amns-2024-1236

SchlüsselwörterGrayscale stretching, Multi-step Markov clustering, Cross-domain image recognition, PGN-RM, Visual design

© 2024 Yi Zhang, published by Sciendo

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

Schlüsselwörter
Grayscale stretching, Multi-step Markov clustering, Cross-domain image recognition, PGN-RM, Visual design