You are on page 1of 9

B. P.

Poddar Institute of Management and Technology

Speech Emotion Recognition


Presented to you by:
Group: 07
Supervised by: SOUMYADEEP MANDAL-11501619006
Prof. T. K. Basu (BPPIMT) ARKADYUTI SARKAR- 11501619024

Dr. Sutapa Mukherjee (BPPIMT) DEBARGHO THANDER- 11501619030


KRISHNENDU CHATEERJEE- 11501619042
Dr. Jayanta Basu (C-DAC)
SUKANYA SADHU- 11501620024
Contents:

• Introduction
• Data Collection
• Data Evaluation
• Feature Extraction
• Spectrogram
• Time Activity Chart
Introduction

“  Speech Emotion Recognition, abbreviated as SER, is the act of


attempting to recognize human emotion and affective states from
speech. This is capitalizing on the fact that voice often reflects
underlying emotion through tone and pitch.

Data Collection

Text
 Determining proper sentences
possessing various emotions used in our
daily lives and checking them by our
supervisors to finalize and keep a record
as digital file.

Speech
Recording of those finalized sentences
by ourselves and also by our targeted
members to use them as voice samples
to extract features and put them in use.
Data Evaluation (Subject Evaluation)

To make our targeted subjects(people) listen both train data & our voice samples (test data) and
take feedback from them on how much they think the voice relates to the mentioned emotions.

Data
Train Data Test Data
 Those are the voice samples for
 Thoseare the predefined voice
samples possessing accurate which feedbacks should be taken
emotions as mentioned. about.
Feature Extraction

 Features like:
• Formant frequencies(,,,)
• Zero-Crossing rate (ZCR)
• Mel-frequency cepstral coefficients (MFCC)

need to be extracted both from train and test data samples for further
processes.
An original Spectrogram of a voice sample possessing surprise
emotion showing different formant frequencies.
Time-Activity Chart
Thank You.

You might also like