Miniature thanka image generation method based on diffusion model

By segmenting and adjusting the layout of the initial image of the miniature Thangka, the problem of layout imbalance caused by inaccurate proportions of image elements in the miniature Thangka was solved. The resulting image layout is more harmonious, the density adjustment of symbolic pattern areas is more accurate, and the style is more distinctive.

CN120411281BActive Publication Date: 2026-06-30TIBET UNIV
View PDF 3 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
TIBET UNIV
Filing Date
2025-04-10
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

In existing technologies, the miniature Thangka image generation model cannot accurately understand the spatial relationship between the subject and the symbolic pattern, resulting in element misalignment or proportional imbalance, and causing an unbalanced image layout.

Method used

By segmenting the initial image of the miniature Thangka, the layout of the main image area and the symbolic pattern area are adjusted separately. Correlation data and proportion adjustment balance index are obtained, style classification is determined, and finally an optimized image that conforms to the style of miniature Thangka is generated.

Benefits of technology

It improves the layout coordination and element proportion accuracy of miniature Thangka images, resulting in more distinctive image styles and more accurate density adjustment of symbolic pattern areas.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120411281B_ABST
    Figure CN120411281B_ABST
Patent Text Reader

Abstract

This invention discloses a method for generating miniature Thangka images based on a diffusion model, belonging to the field of image data processing technology. The method includes the following steps: initial image segmentation of the miniature Thangka, layout optimization of the miniature Thangka, and style determination of the miniature Thangka image. This invention segments the initial image of the miniature Thangka obtained by the diffusion model to obtain a main image region and a symbolic pattern region. Then, it adjusts the main image region's layout to obtain an optimized main region, and adjusts the symbolic pattern region's layout to obtain an optimized symbolic pattern region. Finally, it combines the optimized main region and the optimized symbolic pattern region to obtain an optimized miniature Thangka image, and performs style classification determination. This achieves the effect of improving the layout coordination of the generated miniature Thangka image, solving the problem of inaccurate element proportions leading to unbalanced miniature Thangka image layout in existing technologies.
Need to check novelty before this filing date? Find Prior Art