Study on the development strategy of information integration of packaging design and pattern design under visual communication course

Abstract

In this paper, the importance of modern packaging design and pattern design and the application of pattern design in modern packaging design are the first two aspects of achieving the development of information integration of packaging design and pattern design, and the visual design based on Markov random field is proposed for the image retrieval and classification problems in traditional model packaging design and pattern design. Wavelet transforms, multi-scale analysis of image texture features and region merging conditions in visual design using GBIS algorithm, and image retrieval in pattern design are examined and analyzed. The results show that the model trained with both loss functions has the best retrieval top-k accuracy on both datasets, reaching 54.3% on the DARN dataset and 3.9% accuracy on Deep Fashion. This paper demonstrates the model’s effectiveness for modifying and retrieving pattern design image attributes.

Keywords

• gbis algorithm
• markov random field
• image texture features
• loss function
• package design
• pattern design
• 65d17