Detection of real-time Facial Emotions via Deep Convolution Neural Network
| Autor: | Aysha Rafeeq, S Devipriya, S. Mohana Gowri |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Facial expression
Artificial neural network business.industry Computer science Feature extraction Pattern recognition 02 engineering and technology Overfitting 01 natural sciences Convolutional neural network Convolution Data modeling 0103 physical sciences Classifier (linguistics) 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Artificial intelligence 010306 general physics business |
| Zdroj: | 2021 5th International Conference on Intelligent Computing and Control Systems (ICICCS). |
| DOI: | 10.1109/iciccs51141.2021.9432242 |
| Popis: | Emotions are one of the most prominent features on the human face that let's one study the feelings that another person experiences. It is a way to communicate in a non-verbal way, which is easy to understand, but quite challenging for automated machines. Facial emotion recognition is one of the field of artificial intelligence that has been most promptly researched. This paper demonstrates the classification of realtime human facial expressions into one of the seven categories of emotions using a simple 4-layer Convolution Neural Network(CNN) architecture. The facial features from the input images are extracted and pre-processed through a series of filters and then given to the classifier for the final output. The main goal is to improve the accuracy of the system by deploying various pre-processing and feature extraction techniques to the input images. In addition, data augmentation is done to tackle the overfitting issues. The dataset used for testing and training is FER2013 and the proposed model gives an accuracy of 73%. The FER model can be very useful for business promotions, lie detection, and in areas requiring additional security. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|