Facial expression recognition using deep learning

Facial expression recognition has become an increasingly important area of research in recent years. Neural network- based methods have made amazing progress performing recognition-based tasks, winning competitions set up by various data science communities, and achieving high performance on many datasets. Miscellaneous regularization been utilized researchers to help combat over-fitting, reduce training time, generalize their models. In this paper, applying the Haar Cascade classifier crop faces focus region interest, we hypothesize that would attain a fast convergence without using whole image analyze facial expressions. We also apply label smoothing its effect databases CK+, KDEF, RAF. The ResNet model employed as example neural network model. Label demonstrated improvement accuracy 0.5% considering CK+ KDEF databases. While application shown decrease achieved RAF with small margin, observed.