Deep Learning based Emotion Recognition System Using Speech Features and Transcriptions
This work addresses emotion recognition from speech, which is important for applications like human-computer interaction, but it is incremental as it builds on existing deep learning methods with multimodal inputs.
The paper tackled speech emotion recognition by combining speech features (Spectrogram, MFCC) and text transcriptions, achieving higher accuracies than state-of-the-art methods on the IEMOCAP dataset, with the MFCC-Text CNN model being the most accurate.
This paper proposes a speech emotion recognition method based on speech features and speech transcriptions (text). Speech features such as Spectrogram and Mel-frequency Cepstral Coefficients (MFCC) help retain emotion-related low-level characteristics in speech whereas text helps capture semantic meaning, both of which help in different aspects of emotion detection. We experimented with several Deep Neural Network (DNN) architectures, which take in different combinations of speech features and text as inputs. The proposed network architectures achieve higher accuracies when compared to state-of-the-art methods on a benchmark dataset. The combined MFCC-Text Convolutional Neural Network (CNN) model proved to be the most accurate in recognizing emotions in IEMOCAP data.