Facial Emotion Recognition by Deep CNN and HAAR Cascade
Rohan Shukla1, Agilandeeswari L2, Prabukumar M3
1Rohan Shukla*, School of Information Technology and Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India.
2Agilandeeswari L, School of Information Technology and Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India.
3Prabukumar M, School of Information Technology and Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India.
Manuscript received on September 16, 2019. | Revised Manuscript received on 24 September, 2019. | Manuscript published on October 10, 2019. | PP: 3433-3441 | Volume-8 Issue-12, October 2019. | Retrieval Number: L25891081219/2019©BEIESP | DOI: 10.35940/ijitee.L2589.1081219
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Emotion Recognition is of significance in the modern scenario. Among the many ways to perform it, one of them is through facial expression detection since it is a spontaneous arousal of mental state rather than a conscious effort. Sometimes emotions rule us in the form of the choices, actions and perceptions which are in turn, a result of the emotions we are overpowered by. Happiness, sadness, fear, disgust, anger, neutral and surprise are the seven basic emotions expressed by a human most frequently. In this period of automation and human computer interaction, it is a very difficult and tedious job to make the machines detect the emotions. Facial expressions are the medium through which emotions are shown. For one to detect the facial expression of a person, colour, orientation, lighting and posture play significant importance. Hence, the movements associated with eye, nose, lips etc. plays major role in differentiating the facial features. These facial features are then classified and compared through the trained data. In this paper, we have constructed a Convolution Neural Network (CNN) model and then recognised different emotions for a particular dataset. We have found the accuracy of the model and our main aim is to minimise the loss. We have made use of Adam’s optimizer and used loss function as sparse categorical crossentropy and activation function as softmax. The results which we have got are quite accurate and can be used for further research in this field.
Keywords: Face Detection, HAAR Classifier, Deep Convolutional Network, Cross Validation, Classification- Adam Optimizer, Emotion Recognition.
Scope of the Article: Classification