Abstract : In this paper, we propose a sketch-based interface for wrinkle generation that integrates pen pressure, tilt input, and a skin-muscle deformation model. Existing approaches are limited to static texture-based rendering or linear sketch interpretations using symmetric kernels, which restrict their ability to capture physically plausible wrinkle behavior and user control. To address this, our method transforms freehand sketch inputs into spatially diffused wrinkle shapes using a kernel-based framework. The system incorporates asymmetric kernels and mobile device sensors (pressure and tilt) to more accurately reflect user intent and local variation in wrinkle structure. In addition, we introduce an automatic skin texture shading mechanism that simulates shadow effects around wrinkles and an algorithm that generates fine wrinkles based on user-defined parameters such as age. The interface is implemented via real-time UDP communication with a smartphone, enabling immediate visual feedback and an interactive authoring experience. Experimental results on various facial regions and materials demonstrate that our method produces visually rich and anatomically coherent wrinkles. Quantitative evaluation using the Structural Similarity Index (SSIM) shows that our approach outperforms traditional symmetric kernel-based methods. The proposed system has strong potential for application in digital human creation, forensic sketching, and virtual content design.

[paper]