Facial Expression Recognition Using CNN


  • T. Pradeep Kumar
  • M. Raja Suguna


Facial expressions is non-verbal way of communication among humans through face-face interaction. In recent decades Automatic Expression Detection in an Image or a Video plays an important role due to their drastic application such as Detection of emotions in Psychological Analysis, Cyber Forensics, Emotions Detection in Streaming or stored video for behavioral Analysis. Machine learning Algorithm other than the Neural Network requires complex Feature Extraction Phase followed by the Classification of Emotions, whereas in Classification the accuracy of the model is also less. Traditional Feature Extraction during training phase is also more time consuming. To overcome the difficulties of traditional Approach, Deep Learning Approach called Convolution Neural Network is used to Detection Facial Expression. In Convolution Neural Network, the feature extraction is done by training the Network with the large Facial Expression Image Dataset. The Accuracy of Classification also outperforms than the traditional Machine Learning Algorithms and reaches the States of Art method. Kaggle Facial recognitionFERC-2013 dataset is used.