Robust Emotion Recognition using Sentence, Word and Syllable Level Prosodic Features - Robust Emotion Recognition Using Spectral and Prosodic Features - page 55

Table 3.6 English transcriptions of the Telugu text prompts of IITKGP-SESC
Sentence identity Text prompts

S1 thallidhandrulanu gauravincha valenu

S2 mI kOsam chAlA sEpatnimchi chUsthunnAmu

S3 samAjamlo prathi okkaru chadhuvuko valenu

S4 ellappudu sathyamune paluka valenu

S5 I rOju nEnu tenali vellu chunnAnu

S6 kOpamunu vIdi sahanamunu pAtincha valenu

S7 anni dAnamulalo vidyA dAnamu minnA

S8 uchitha salahAlu ivvarAdhu

S9 dongathanamu cheyutA nEramu

S10 I rOju vAthAvaranamu podigA undhi

S11 dEsa vAsulandharu samaikhyAthA tho melaga valenu

S12 mana rAshtra rAjadhAni hyderAbAd

S13 sangha vidhrOha sekthulaku Ashrayam kalpincharAdhu

S14 thelupu rangu shAnthiki chihnamu

S15 gangA jalamu pavithra mainadhi

the syllable level, is fixed at 4, which is equal to the maximum number of syllables
in any group, as shown in Table 3.5 .
Out of 10 speakers of IITKGP-SESC, the speech utterances of eight speakers
(4 male and 4 female) are used to train the emotion recognition models. Validation
of the trained models is done using remaining 2 speakers' (1 male and 1 female)
speech data. The details of the speech corpus, IITKGP-SESC, are given in Sect. 2.2
of Chap. 2 . The description of development of emotion recognition models and their
verification is discussed in the next Section.

3.5 Results and Discussion

Emotion recognition systems are separately developed for sentence, word, and
syllable level global and local level prosodic features. The combination of global
and local level features is also explored to study emotion recognition (ER).

3.5.1 Emotion Recognition Systems using Sentence Level
Prosodic Features

In this work, we have considered 8 emotions of IITKGP-SESC, for studying the
role of global and local prosodic features in recognizing speech emotions. SVMs are
used to develop emotion recognition models. Each SVM is trained with positive and