Deep Learning Method and Apparatus for Emotion Recognition based on Efficient Multimodal Feature Groups and Model Selection

摘要

A deep learning method and apparatus for effective model and feature group selection in emotion recognition using various Asian databases are presented. The deep learning apparatus for effective model and feature group selection in emotion recognition using various databases of Asians proposed in the present invention is EEG ( A feature extracting unit that extracts electroencephalogram) features, an LSTM model selector that selects an LSTM model to apply to the extracted EEG features using a genetic algorithm (GA), and a genetic algorithm for the selected LSTM model. and a feature set selector for selecting a set.