Facial Expression Recognition Based on Local Features of Transfer Learning
Haiqiang Feng, Jingfeng Shao · 2020 IEEE 4th Information Technology, Networking, Electronic and Automation Control Conference (ITNEC) · 2020
Due to the difficulty and low accuracy of facial expression recognition caused by facial occlusion issue, a local-face expression recognition model based on migration learning was proposed and Inception-v3 pre-training model was used to build the network. Histogram equalization, ROF denoising, affine change, image correction and other image preprocessing were carried out on the image data. Extract the data characteristics of preprocessed data by using the Inception-v3 model, and put the extracted feature information into a new classifier for classification; The accuracy of the proposed model was tested by CK+ and Jaffe data sets, and compared with other models, the accuracy was up to 98.2%, which proved that the proposed model had strong robustness for facial expression recognition of face data covered by the face.