Optimized synthesis of art patterns and layered textures

Area of Science:

• Computer Graphics
• Digital Art
• Image Synthesis

Background:

• Art patterns and textures are prevalent in digital media, from icons to illustrations.
• Existing exemplar-based texture synthesis methods have shown success.
• Textures with discrete, curved elements share characteristics with art patterns.

Purpose of the Study:

• To develop a method for synthesizing art patterns and textures with curvilinear features from exemplars.
• To treat art pattern synthesis as a global optimization problem.
• To extend the method for multilayer pattern and texture synthesis.

Main Methods:

• Formulated an energy function measuring color appearance and curvilinear shape similarity.
• Developed an expectation-maximization-style algorithm for energy minimization.
• Applied the level set method to optimize shape similarity.

Main Results:

• Successfully synthesized art patterns and textures with curvilinear features from input exemplars.
• The optimization algorithm effectively minimized both color and shape similarity metrics.
• The generalized approach handled multilayer synthesis and interactions.

Conclusions:

• The proposed method provides an effective approach to synthesizing complex art patterns and textures.
• The integration of level sets offers an innovative solution for shape similarity in synthesis.
• The generalized algorithm supports multilayer synthesis, expanding its applicability.