摘要
通过梳理当前国内外人工智能生成字库的方法,比较分析不同生成方法的优势与劣势,选择图像转换技术下的人工智能学习模型作为技术支持。以盲文汉字字体作为实验对象,结合此字体结构突破常规、笔画平直等特点,将笔画组字方法论引入人工智能,提出先"部件"再"整字"两步式的创新方法进行设计实验,最终实现盲文汉字字库的自动化生成。实验表明两步式的创新生成方法能进一步提高计算机生成字库的精确性与识别度,使盲文汉字可以高效转变为标准化的字库产品,也为自动化生成风格一致、字型准确、符合规范的个性化字库产品提供一定理论依据。
This paper studies how to use artificial intelligence to expand Braille Chinese characters in order to provide certain theoretical basis for automatic generation of font products with consistent style and standard.This paper combs the current methods of artificial intelligence generation font at home and abroad,selects the suitable training model,takes Braille Chinese character font as the experimental object,and proposes a two-step artificial intelligence generation idea of"part"and"whole character"to carry out design experiment,so as to realize the automatic generation of Braille Chinese character library.The experimental results show that the two-step innovative method can further improve the accuracy and recognition of computergenerated font.
作者
陆赉
关辛
LU Lai;GUAN Xin(University of Shanghai for Science and Technology,200093 Shanghai;East China University of Science and Technology,200237 Shanghai)
出处
《设计艺术研究》
2021年第6期49-52,共4页
Design Research
基金
国家社科基金艺术学项目(2019CG04337)
2020年度上海理工大学教师教学发展研究项目(CFTD202003)。
关键词
人工智能
盲文汉字
GANs
字库生成
Artificial Intelligence
Braille Chinese characters
GANs
font generation