摘要
Facial expression recognition is a hot topic in computer vision, but it remains challenging due to the feature inconsistency caused by person-specific 'characteristics of facial expressions. To address such a challenge, and inspired by the recent success of deep identity network (DeepID-Net) for face identification, this paper proposes a novel deep learning based framework for recognising human expressions with facial images. Compared to the existing deep learning methods, our proposed framework, which is based on multi-scale global images and local facial patches, can significantly achieve a better performance on facial expression recognition. Finally, we verify the effectiveness of our proposed framework through experiments on the public benchmarking datasets JAFFE and extended Cohn-Kanade (CK+).
Facial expression recognition is a hot topic in computer vision, but it remains challenging due to the feature inconsistency caused by person-specific 'characteristics of facial expressions. To address such a challenge, and inspired by the recent success of deep identity network (DeepID-Net) for face identification, this paper proposes a novel deep learning based framework for recognising human expressions with facial images. Compared to the existing deep learning methods, our proposed framework, which is based on multi-scale global images and local facial patches, can significantly achieve a better performance on facial expression recognition. Finally, we verify the effectiveness of our proposed framework through experiments on the public benchmarking datasets JAFFE and extended Cohn-Kanade (CK+).
基金
supported by the Academy of Finland(267581)
the D2I SHOK Project from Digile Oy as well as Nokia Technologies(Tampere,Finland)
作者简介
Chongsheng Zhang was born in 1982. He is a fullprofessor of Henan University, China, where he isalso the director of the big data research. He receivedhis Ph.D. degree at INRIA, France. He haspublished more than 20 papers in peer-reviewedconferences and journals, including IEEE ICDM,PAKDD. He has (co-) authored three books andholds three patents. His research interests includedata classification, data stream mining and deep learning.E-mail: Chongsheng.zhang@yahoo.com;Pengyou Wang was born in 1990. He is currentlya second-year master student at Henan University,China. He has lead or participated in many projects,including knowledge graph based intelligent jobmatching system, property management system, facialexpression recognition and insect sound recognition.He has co-authored one book which is aboutdeep learning and the usage of Caffe. His researchinterests are deep learning and pattern recognition.E-mail: 1204287950@qq.com;Corresponding author.Ke Chen was born in 1985. He received his Ph.Ddegree in computer vision at the School of ElectronicEngineering and Computer Science, QueenMary, University of London, UK. He is currently theAcademy of Finland post-doctoral research fellowat the Department of Signal Processing, TampereUniversity of Technology. His research interests includecomputer vision, pattern recognition, neuraldynamic modelling, and robotic inverse kinematics.E-mail: ke.chen@tut.fi;Joni-Kristian Kamarainen was born in 1974. Heis an associate professor of signal processing at theDepartment of Signal Processing, Tampere Universityof Technology, Finland. He received his M.S.and Ph.D. degrees from Lappeenranta University ofTechnology in 1999 and 2003, respectively. He leadsthe Computer Vision Group and his research focuseson 2D and 3D scene analysis, object detection andrecognition, signal processing and machine intelligence.E-mail: joni.kamarainen@tut.fi
