Novel approach to emotion recognition in voice: a convolutional neural network approach and grad-cam generation

Download

index.pdf

Date

2019

Author

Canpolat, Salih Fıra

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

192
views

145
downloads

Emotion is one of the essential components in human and human-machine interaction. One of the most common communication channels is the sound. Understanding the underlying mechanisms of emotion recognition in the sound signal is an essential step in improving both types of interaction. For this purpose, we developed an emotion recognition model, and a Turkish-specific database, referred to as the Turkish Emotion-Voice (TurEV) database. The database contains one-word-vocalizations of four emotion types; angry, calm, happy, and sad in three different frequency bands. The model was trained using TurEV, and human validation studies were conducted. The results indicate that the model is feasible for emotion recognition tasks. The comparison of the humans with the computational model indicate that the model achieves better results using feature-rich frequency bands, the humans use all other aspects of the sound signal.

Subject Keywords

Emotions., cnn, emotion, voice, corpus, Turkish

URI

http://etd.lib.metu.edu.tr/upload/12623618/index.pdf
https://hdl.handle.net/11511/43913

Collections

Graduate School of Informatics, Thesis

Suggestions

OpenMETU
Core

Predictors of components of expressed emotion in major caregivers of Turkish patients with schizophrenia Karancı, Ayşe Nuray (2002-02-01) Background Expressed emotion (EE) is a concept reflecting the emotional atmosphere of the home environment. Specific components of EE, namely criticism, hostility and emotional over-involvement, have been found to be important predictors of relapse for schizophrenic patients. The main aim of this study was to examine the predictive power of patient and caregiver characteristics and caregivers' perceptions of frequency, coping, distress/discomfort, control of symptom behaviours by the patient, and attributio...
Examination of Computer Supported Collaborative Business Process Modeling with Activity Theory FINDIK COŞKUNÇAY, DUYGU; Çakır, Murat Perit (2014-09-10) Activity Theory provides a framework to examine and explain human-human and human-computer interactions. In this study, Activity Theory was used to examine both interactions in depth in the context of Computer Supported Collaborative Business Process Modeling (CSCBPM) in which geographically dispersed multiple users interact with each other and with the system. This framework enabled us to examine the activities of CSCBPM in detail and understand the process of CSCBPM. During the CSCBPM, some difficulties w...
Learning to Generate Unambiguous Spatial Referring Expressions for Real-World Environments Dagan, Fethiye Irmak; Kalkan, Sinan; Leite, Iolanda (2019-01-01) Referring to objects in a natural and unambiguous manner is crucial for effective human-robot interaction. Previous research on learning-based referring expressions has focused primarily on comprehension tasks, while generating referring expressions is still mostly limited to rule-based methods. In this work, we propose a two-stage approach that relies on deep learning for estimating spatial relations to describe an object naturally and unambiguously with a referring expression. We compare our method to the...
Influence of Aesthetic Properties on Stimulating Emotional Responses Sevener, Zeynep; Asatekin, Mehmet (2019-4-1) The purpose of this paper is to provide a framework that demonstrates the role of aesthetic properties in stimulating emotional experiences. The framework is constructed as the answer to the question: "What are the consequences of the stimulus of aesthetic properties on product related emotions and experiences?”. The focus of the study is in investigating the links between the visual qualities of the products and the emotional experiences. The immediate sensorial experiences transpire during the initial ste...
Self-organised Flocking of Robotic Swarm in Cluttered Environments Liu, Zheyu; Turgut, Ali Emre; Lennox, Barry; Arvin, Farshad (2021-01-01) Self-organised flocking behaviour, an emergent collective motion, appears in various physical and biological systems. It has been widely utilised to guide the swarm robotic system in different applications. In this paper, we developed a self-organised flocking mechanism for the homogeneous robotic swarm, which can achieve the collective motion with obstacle avoidance in a cluttered environment. The proposed mechanism introduces an obstacle avoidance approach to the Active Elastic Sheet model that was previo...

Citation Formats

S. F. Canpolat, “Novel approach to emotion recognition in voice: a convolutional neural network approach and grad-cam generation,” Thesis (M.S.) -- Graduate School of Informatics. Cognitive Sciences., Middle East Technical University, 2019.