Example-Based Floral Pattern Generation

An example-based floral pattern generation approach is proposed in this paper to improve the aesthetic design ability of computer, which will enable user’s simple draft to be converted to an aesthetic pattern with a few interactions. The proposed approach commences with introducing artists work into the system by example patterns they drew. User’s draft is segmented into elements, shape features of element are extracted and matched with that of stylized elements in database, then the constraints in draft are used to regulate the matched stylized elements and generate a new pattern. Moreover, we proposed a novel weighted similarity measure for the purpose of improving matching accuracy. Results show that artistic pattern can be generated from simple draft with our method.