| Summary: | 碩士 === 銘傳大學 === 資訊傳播工程學系碩士班 === 105 === The aging society with fewer children is coming recently. Elderly population will be the main influence to the society, such as loneliness among the elderly. In the meantime, if we have a companion robot, which will help and increase quality of caring with the elderly, it can reduce many society issues. This study presents The Research on Human Emotion Recognition Based on Deep Learning Algorithm. With the deep learning algorithm, this research can reorganize and predict seven different kinds of emotions. These emotions include happiness, sadness, surprise, anger, disgust, fear and neutral. The system uses the Deep Convolution Neural Network to train and produce the prediction. While using SVM and SPSD for pre-process the facial of human, this system achieves about 89.2% of accuracy on predicting the emotions. This high accuracy helps the machine to learn emotion human much easier and maybe apply on the robot for other application. This research can be applied on the companion robot and helps the robots become more humanized in the future.
|
|---|
