@Article{JFBI-16-257,
author = {Shao , Hong-En},
title = {Application of Semantic Image Generation Techniques Based Detail Preserving Image Method for Repoussé Craft},
journal = {Journal of Fiber Bioengineering and Informatics},
year = {2024},
volume = {16},
number = {3},
pages = {257--267},
abstract = {<p style="text-align: justify;">Text-to-image technology is a technology for mapping target image sets through natural language and
is the latest production method for new content. The repoussé craft1
is ancient Chinese gold and silver
fine art crafts. Although the craft has been studied, its exploration in digital research is relatively new.
To modernise the repoussé craft, this paper takes the combination of 3D generated drawings of repoussé craft motifs with text-to-image technology as a research object, applies current artificial intelligence
to the study and analyses the possibilities of the combined application and related text specification
recommendations, tests and analyses the impact of the changes in the prompt and weighted value
settings on the repoussé craft jewelry design style. The results of the study show that the way the
prompt2 and matting image weights are set plays a key role in the stabilisation of the imaging style.
This study not only fills the gap in the combination of repoussé technology and AI technology but also
plays a positive role in the stability of the style of the enhanced text to imaging, which provides a specific
parameter basis and a new way of thinking for the future application of AI technology in the design of
fashion jewelry and even in the wider field of fashion design.</p>},
issn = {2617-8699},
doi = {https://doi.org/10.3993/jfbim02322},
url = {http://global-sci.org/intro/article_detail/jfbi/23081.html}
}