Details

Autor(en) / Beteiligte

• Zheng, Dongyi
• Wang, Haotian
• Liu, Panyu
• Liu, Fang
• Xiao, Nong

Titel

GDAssister: Graphic Design Assistant System with Optimal Algorithm of Associated Rule Mining

Ist Teil von

• Theoretical Computer Science, p.57-69

Ort / Verlag

Singapore: Springer Singapore

Link zum Volltext

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• In the process of modern graphic design, designers usually get inspiration from a large number of art graphs and then choose graphic elements for graphic design. However, the process is generally manual, limiting the alternatives a designer can explore. In this paper, we propose GDAssister, a graphic design assistant system with optimal algorithm of associated rule mining. The system can automatically generate corresponding graphs according to the requirements of designers and use style transfer technology to change the style (color and texture) of the graphs. The system introduces four modules referring to the process of graphic design (e.g., graphic element extraction, association rule mining, automatic layout, style transform). Association rule mining is the core module of the system, and the efficiency of association rule mining algorithm determines the running efficiency of GDAssister. We design an improved association rule mining algorithm (IFP-Growth), which can reduce time consumption and memory usage. We use Chinese ethnic apparel database to verify the feasibility and efficiency of the system. Our experiment shows that GDAssister: (1)helps designers create more graphic alternatives to explore design creativity and (2)helps designers generate different styles of graphs more quickly to shorten the design cycle.