| 基于语义维度的人脸表情生成 |
| |
| 引用本文： | 张申,贾珈,王晓慧,蔡莲红.基于语义维度的人脸表情生成[J].清华大学学报(自然科学版),2011(1):80-84. |
| |
| 作者姓名： | 张申 贾珈 王晓慧 蔡莲红 |
| |
| 作者单位： | 清华大学计算机科学与技术系普适计算教育部重点实验室清华信息科学与技术国家实验室; |
| |
| 基金项目： | 国家“九七三”重点基础研究项目(2011CB302201); 国家自然科学基金资助项目(90820304,60805008) |
| |
| 摘 要： | 为实现有丰富表现力的人脸表情生成,提出一种基于语义维度的人脸表情生成方法。首先定义了与人类情感、态度、意图等情态信息相关的7个语义维度,对表情图像进行语义维度标注;进而对面部器官的典型状态进行量化,定义并提取人脸局部状态参数。分别采用线性回归和人工神经网络方法,建立了语义维度与人脸局部状态参数的映射模型。实验结果表明:基于语义维度的方法能够生成与文本语义相贴切的人脸表情,神经网络模型在参数预测精度和主观感知评价方面均优于线性模型。
|
| 关 键 词： | 人脸表情 语义维度 人工神经网络 线性回归 |
Facial expression synthesis based on semantic dimensions |
| |
| ZHANG Shen,JIA Jia,WANG Xiaohui,CAI Lianhong.Facial expression synthesis based on semantic dimensions[J].Journal of Tsinghua University(Science and Technology),2011(1):80-84. |
| |
| Authors: | ZHANG Shen JIA Jia WANG Xiaohui CAI Lianhong |
| |
| Institution: | ZHANG Shen,JIA Jia,WANG Xiaohui,CAI Lianhong(Key Laboratory of Pervasive Computing of Ministry of Education,Tsinghua National Laboratory for Information Science and Technology,Department of Computer Science and Technology,Tsinghua University,Beijing 100084,China) |
| |
| Abstract: | An approach was developed to synthesize facial expressions using semantic dimensions to synthesize facial expressions with rich expressivity.Sever semantic dimensions were defined to describe the modality information such as emotion,attitude and intentions,with these semantic dimensions used to annotate the facial expression images.Partial facial parameters were defined to parameterize typical states of the facial features.The semantic dimensions were mapped to the partial facial parameters based on a linea... |
| |
| Keywords: | facial expression semantic dimension artificial neural network linear regression |
| 本文献已被 CNKI 等数据库收录！ |
|
