Facial Expression Recognition By Fusing Local Facial Regions
Ying Zi-lu, Youwei Zhang, Jingwen Li · Signal Processing · 2009
A novel method for facial expression recognition by fusing local facial regions is proposed.The new approach first extracts local regions such as mouth,nose and eyes.Then expression recognition is carried out based on these regions as well as on the whole face.The final facial expression recognition result is obtained by fusing recognition results based on local facial regions.The proposed approach is experimented extensively on both Japanese female facial expression database JAFFE and Chinese Expression Database Wuyi university facial expression database CED_WYU.The experiment results show that the facial expression recognition rates increase obviously compared with those based only on local regions or only the whole face and prove the effectiveness of the proposed fusion approach.