《基于样本的人脸线条画非参数化生成》中学生教育文学丛书.pdfVIP

Example-based Facial Sketch Generation with Non-parametric Sampling 1  1 1 2 3 Hong Chen , Ying-Qing Xu , Heung-Yeung Shum , Song-Chun Zhu and Nan-Ning Zheng 1 Microsoft Research, China 2 The Ohio State University, USA 3 Xi’an Jiaotong University, China Abstract caricature system PICASSO and Web-PICASSO[14]. Li et al.[8] proposed an automatic facial sketch system that uses a In this paper, we present an example-based facial sketch generalized symmetry operator, rectangle filter, and charac- system. Our system automatically generates a sketch from teristic shapes to detect the locations of facial feature points. an input image, by learning from example sketches drawn Many approaches have also been proposed to generate fa- with a particular style by an artist. There are two key el- cial caricature [10, 6]. But these approaches, without ob- ements in our system: a non-parametric sampling method serving and learning from the artist’s products, in general and a flexible sketch model. Given an input image pixel and produce stiff and inexpressive sketches. its neighborhood, the conditional distribution of a sketch point is computed by querying the examples and finding In fact, few systems have been proposed to teach a com- all similar neighborhoods. An “expected sketch image” is