Robust Emotion Recognition using Pitch Synchronous and Sub-syllabic Spectral Features - Robust Emotion Recognition Using Spectral and Prosodic Features - page 42

Table 2.10 continued

Features Methods & IIT Emo-

regions for KGP- DB

feature extr. SESC

LPCCs Consonant 44 40

MFCCs region 44 42

LPCCs + Formants 51 51

MFCCs + Formants 36 34

LPCCs CV- 65 66

MFCCs transition 58 57

LPCCs + Formants region 71 69

MFCCs + Formants 68 64

LPCCs Pitch 69 67

MFCCs synchronous 68 65

LPCCs + Formants analysis 73 72

MFCCs + Formants 72 70

Table 2.11 Average emotion classification performance using proposed spectral features on Set1,
Set2, and Set3 datasets of IITKGP-SESC
Features Methods & IITKGP-SESC (rec.%)

regions for Set1 Set2 Set3

feature extr.

LPCCs Block 74 71 69

MFCCs processing 67 65 65

Formant features (Entire 53 49 47

LPCCs + Formants speech 75 72 70

MFCCs + Formants Signal) 71 68 67

LPCCs Vowel 57 55 54

MFCCs region 56 54 51

LPCCs + Formants 65 62 60

MFCCs + Formants 53 52 51

LPCCs Consonant 47 45 44

MFCCs region 47 44 44

LPCCs + Formants 55 53 51

MFCCs + Formants 40 38 36

LPCCs CV- 68 66 65

MFCCs transition 63 60 58

LPCCs + Formants region 74 72 71

MFCCs + Formants 72 71 68

LPCCs Pitch 75 73 69

MFCCs synchronous 72 69 68

LPCCs + Formants analysis 77 75 73

MFCCs + Formants 74 73 72