Research On Real-life Facial Expression Recognition Based On Deep Learning

Facial expression recognition is a research hotspot in deep learning.The large intra-class gap seriously affects the accuracy of facial expression recognition.How to use facial recognition to improve the accuracy of security checks has become the top priority for ensuring environmental safety.Therefore,solving the problem of large intra-class gaps in facial expression recognition,accurately identifying facial expressions,and identifying suspicious molecules being monitored are particularly important for solving complex environmental safety issues.Although facial recognition intelligent recognition technology has a long history of research,the technology is relatively mature,but because of the large gap in facial expressions,the gap between the classes is small,the external influence factors are more,and the recognition results are not satisfactory.Deep learning,because of its powerful expressive ability,has become a research hotspot in the field of facial expression recognition in real environment.This paper conducts further research on real-life facial expression recognition based on deep learning.The specific work is as follows:(1)This paper expounds the research status of facial expression recognition at home and abroad,introduces the model structure of traditional neural network,and carries out corresponding contrast experiments and analysis under the background of real environment facial expression recognition,and proposed based on the depth of learning IC-GAN(Intral-Class Gap GAN)real environment facial expression recognition method;(2)The sample data is obtained by means of network downloading and video.For the problem that the data has a large intra-class difference,the data set is pre-processed by image expansion and the like,and the complexity of facial expression features with a large number of intra-class differences is added.Improved accuracy of expression recognition provides a reliable data foundation;(3)The constructed IC-GAN network consists of a convolutional layer,a fully connected layer,an active layer,a BN(BathNorm)layer,and a Softmax layer.The convolutional encoder and decoder are used to perform deeper feature extraction on the facial expression image to ensure that the feature is extracted.The accuracy of facial expression recognition;(4)Both network training and network identification use momentum-based Adam to update network weights,adjust network parameters to optimize network structure,reduce network error in facial expression recognition,and improve recognition accuracy;(5)The Python GUI programming method is used to design the interactive interface,and the recognition result is applied to the actual production life to realize the expression recognition.In order to solve real-world environment recognition rate caused by the large disparities in the class for complex environments subway,railway stations,airports,etc.,to build a new IC-GAN network identification model to make it better adapted to the Complex facial expression recognition tasks.This paper trains the facial expression category data based on the Pytorch platform in deep learning,and tests it on the test set of the self-made mixed facial expression data set.When the input image is 256×256,the recognition accuracy is as high as 84.9%,IC-GAN While ensuring the accuracy,the network model reduces the misrecognition rate of expressions in the case of large gaps in the class,blurred images and incomplete facial expressions,and improves the robustness of the system.