Artistic Font Generation from Character Stroke Extraction and Controllable Image Diffusion Models

In light of the burgeoning integration of artificial intelligence(AI) into the field of graphic design, we present an innovative framework tailored for artistic font generation, specifically aimed at enhancing the creative process of typography design. By Combining Chinese character stroke extraction with controllable image diffusion models, our framework facilitates the generation and manipulation of artistic fonts at both character and stroke levels. Through qualitative and quantitative analyses, our method showcases precise control over the generation of specified character strokes, as well as exceptional fidelity to textual descriptions. Notably, it exhibits remarkable artistic quality in the generated font images, thereby offering a promising avenue for advancing the realm of typographic design.