Data-Driven Synthesis of Cartoon Faces Using Different Styles

Abstract

Abstract-This paper presents a data-driven approach for automatically generating cartoon faces in different styles from a given portrait image. Our stylization pipeline consists of two steps: an offline analysis step to learn about how to select and compose facial components from the databases; a runtime synthesis step to generate the cartoon face by assembling parts from a database of stylized facial components. We propose an optimization framework that, for a given artistic style, simultaneously considers the desired image-cartoon relationships of the facial components and a proper adjustment of the image composition. We measure the similarity between facial components of the input image and our cartoon database via image feature matching, and introduce a probabilistic framework for modeling the relationships between cartoon facial components. We incorporate prior knowledge about image-cartoon relationships and the optimal composition of facial components extracted from a set of cartoon faces to maintain a natural, consistent, and attractive look of the results. We demonstrate generality and robustness of our approach by applying it to a variety of portrait images and compare our output with stylized results created by artists via a comprehensive user study.Index Terms-Cartoon face, face stylization, data-driven synthesis, component-based modeling.Y. Zhang is with the National

Painting Many Pasts: Synthesizing Time Lapse Videos of Paintings

“…In style transfer, images are transformed to simulate a specific style, such as a painting-like style [20,21] or a cartoon-like style [66]. More recently, neural networks have been used for generalized artistic style transfer [18,69].…”

Efficient PatchMatch-Based Synthesis for Cartoon Animation

“…After applying the irregular neighborhoods in temporal domain, work in [6] enables a more general framework for dynamic element textures, from which we draw inspiration to pursue benefits of data-driven approaches for synthesizing dynamic contents in 2D animation. Besides, there are other interesting data driven synthesis methods offering the state-of-the-art results, but most of them aim at specific applications, such as cloth wrinkles [27], crowds [28], faces cartooning [29], [30] and learning basketball dribbling skills [31]. Therefore, we draw our inspiration from all the above investigations to pursue similar benefits of data-driven approaches for synthesizing dynamic contents in 2D animation, which tends to be more general and user friendly.…”

Watercolour Rendering of Portraits

“…In addition, the facial model is used to enhance the rendering of the eyes, lips, teeth and highlights. Zhang et al [27] describe an approach to creating cartoon versions of faces. Facial components are detected in the input image, and are matched to a dictionary of cartoon stylised components.…”