Emotion Analysis and

Recognition Using ECG

and Machine Learning
Avila, John Paul F., Nicomedes, Charlene Dianne D.,
Orioque, Abby Kate C.
Abstract
The electrocardiogram (ECG), which is widely used to detect and
track cardiovascular disease, is created from electrical impulses of
the heart's activities. The extensive availability of wearable devices
that can collect ECG or PPG on the market supports the claim that
signals can be employed in the field of affective computing to
identify human emotions. There are numerous methods for creating
emotion recognition systems today using different methodologies and
algorithms. This review paper focuses on research on emotion
recognition that uses ECGs and AI as a unimodal approach as well as
a component of a multimodal strategy for emotion recognition
systems. The results presented here are helpful for aspiring
researchers to review a summary of earlier studies done in the field of
ECG-based emotion recognition systems, to identify gaps in the field,
and to develop and design future applications of emotion recognition
systems, particularly in improving healthcare.

9/3/20XX Presentation Title 2

 DREAMER: A Public Dataset for
ECG & EEG Emotion Recognition

9/3/20XX Presentation Title 3

9/3/20XX Presentation Title 4
Data Processing Procedures
• 1. Participants are to watch different videos, varying in genre
such as comedy, horror, calm scenes, and other types of scenarios
that invokes different emotions.
• 2. In total, participants will watch 3 minutes worth of videos that
goes through the 6 basic emotions (happy, anger, fear, sad,
surprise, neutral)
• 3. While they are watching data are gathered using an ECG
sensor that is attached to their arms and leg.

9/3/20XX Presentation Title 5

Participants
• Sample Size: 10 Volunteers
• Sex: Male/Female
• Age Group: 6 y.o. – 24 y.o.
• Tools: ECG Sensor, Movie Clips

9/3/20XX Presentation Title 6

The Valence-Arousal-Dominance
Model
Valence: Pleasantness

Arousal: Intensity

Dominance: Degree of Control

9/3/20XX Presentation Title 7

RESULTS AND DISCUSSION

Participant Count: 10 Volunteers

9/3/20XX Presentation Title 8

RESULTS AND DISCUSSION

Participant Count: 10 Volunteers

9/3/20XX Presentation Title 9

Conclusion

In this paper, we presented a feasible method of

emotion recognition based on ECG signal. The
obtained results is better than previous researches in
the emotion detection based on ECG signal. In
addition, according to the classification
consequences, the effectiveness of feature extraction
was proved. The recognition rate approached to the
one based on facial recognition, but the method
based on ECG signal is much simpler on steps and
data processes. In addition, the ECG signals can truly
react the emotion of individuals, because it is
difficult for individuals to mask the change on ECG
signals

9/3/20XX Presentation Title 10

Program Runs

9/3/20XX Presentation Title 11

Program Runs

9/3/20XX Presentation Title 12

Program Runs

9/3/20XX Presentation Title 13

Thank you

9/3/20XX Presentation Title 14