Feature Representation for Speech Emotion Recognition

Feature Representation for Speech Emotion Recognition Publisher

Abdollahpour M¹ ; Zamani J¹ ; Rad HS²

Show Affiliations

1. Department of Electrical Engineering, Amirkabir University of Technology, Tehran Polytechnic, Iran
2. Department of Medical Physics and Biomedical Engineering, Tehran University of Medical Sciences (TUMS), Research Center for Science and Technology in Medicine (RCSTIM), Imam Khomeini Hospital Complex, Tehran, Iran

Source: 2017 25th Iranian Conference on Electrical Engineering# ICEE 2017 Published:2017

Abstract

In this paper we proposed a method for feature representation and enhancement algorithm based on denoising auto encoder (DAE) for speech emotion recognition. We defined new similarity and dissimilarity cost functions that enforce features relating to the same class be similar and features corresponding to different classes be dissimilar. We represented a new supervised DAE architecture that learns features robust to speaker and linguistic variations. Experimental results demonstrate significant improvement over original features. © 2017 IEEE.

Style	Citing Format
MLA	Abdollahpour M, Zamani J, Rad HS. "Feature Representation for Speech Emotion Recognition." 2017 25th Iranian Conference on Electrical Engineering# ICEE 2017, vol. , no. , 2017, pp. 1465-1468.
APA	Abdollahpour M, Zamani J, Rad HS (2017). Feature Representation for Speech Emotion Recognition. 2017 25th Iranian Conference on Electrical Engineering# ICEE 2017, (), 1465-1468.
Chicago	Abdollahpour M, Zamani J, Rad HS. "Feature Representation for Speech Emotion Recognition." 2017 25th Iranian Conference on Electrical Engineering# ICEE 2017 , no. (2017): 1465-1468.
Harvard	Abdollahpour M, Zamani J, Rad HS (2017) 'Feature Representation for Speech Emotion Recognition', 2017 25th Iranian Conference on Electrical Engineering# ICEE 2017, (), pp. 1465-1468.
Vancouver	Abdollahpour M, Zamani J, Rad HS. Feature Representation for Speech Emotion Recognition. 2017 25th Iranian Conference on Electrical Engineering# ICEE 2017. 2017;():1465-1468.
BibTex	@article{ author = {Abdollahpour M and Zamani J and Rad HS}, title = {Feature Representation for Speech Emotion Recognition}, journal = {2017 25th Iranian Conference on Electrical Engineering# ICEE 2017}, volume = {}, number = {}, pages = {1465-1468}, year = {2017} }
RIS	TY - JOUR AU - Abdollahpour M AU - Zamani J AU - Rad HS TI - Feature Representation for Speech Emotion Recognition JO - 2017 25th Iranian Conference on Electrical Engineering# ICEE 2017 VL - IS - SP - 1465 EP - 1468 PY - 2017 ER -

Science Communicator Platform

Authors

Authors Affiliations

Abstract