EMOTION RECOGNITION SYSTEMS:
INSIGHTS FROM BIOLOGY
Mehmet A.
Duru G.
Azra Y.
Ada T.
WHAT ARE EMOTIONS?
“EMOTIONS ARE A PROCESS, A PARTICULAR KIND OF AUTOMATIC
APPRAISAL INFLUENCED BY OUR EVOLUTIONARY AND PERSONAL
PAST, IN WHICH WE SENSE THAT SOMETHING IMPORTANT TO OUR
WELFARE IS OCCURRING , AND A SET OF PSYCHOLOGICAL CHANGES
AND EMOTIONAL BEHAVIORS BEGINS TO DEAL WITH THE SITUATION."
- PAUL EKMAN, PHD
WHY ARE EMOTIONS There are many times, often during periods of great
challenges, that emotions can feel more like a burden than a
IMPORTANT?
blessing.
But emotions can be powerful
They can help you survive, grow, and connect with others.
And they can guide your decisions, behaviors, and
motivations.
As babies, emotions are how you learn to communicate,
even before you can talk.
While intense emotions can feel like a lot, without them, life
can feel bland, muted, and empty. Emotions are an integral
part of being human.
INTRODUCTION
- What are the emotion recognition systems?
Emotion recognition systems are technologies designed to identify and interpret human
emotions based on various data sources, such as:
FACIAL EXPRESSIONS
VOICE TONE
BIOMETRIC SIGNALS (e.g. heart rate, skin conductance)
1. FACIAL EXPRESSIONS
Based on Paul Ekman’s research on universal emotions (Happiness, sadness, anger, fear, surprise, disgust.):
Facial Action Coding System (FACS) will maps muscle movements to emotions.
Analysis Process:
• Step 1: Face detection by using computer cameras.
• Step 2: Facial feature analysis (e.g. eye movements, mouth curvature).
• Step 3: Emotion classification via machine learning.
Emotions
Graph
2. VOICE PARAMETERS
• Pitch (Frequency): A higher pitch often indicates happiness or excitement, while a lower pitch may
reflect sadness or seriousness.
• Volume (Loudness): Louder voices can signify anger or enthusiasm, while softer voices might
indicate sadness or shyness.
• Speech Rate: Faster speech may signal stress or anxiety, while slower speech often reflects
calmness or sadness.
Analysis Process:
• Step 1: Collecting audio data (e.g., via a microphone).
• Step 2: Extracting features like frequency, intensity, rhythm, and tempo.
• Step 3: Classifying these features into emotion categories using machine learning models.
3. BIOMETRIC SIGNALS
Heart Rate Variability (HRV):
• Variations in heart rate reflect the activity of the sympathetic and parasympathetic nervous systems.
• High HRV often indicates relaxation, while low HRV may signal stress.
Galvanic Skin Response (GSR):
• Measures changes in the electrical conductivity of the skin due to sweat gland activity.
• Higher GSR values indicate intense emotions like fear or excitement.
Pupil Dilation:
• Pupil size changes in response to emotional arousal and attention levels.
Analysis Process:
• Step 1: Data collection using sensors (e.g., ECG devices, skin conductance sensors).
• Step 2: Categorising the data’s by their groups.
• Step 3: Classifying emotional states using AI models.
APPLICATIONS OF EMOTION
RECOGNITION SYSTEMS
Emotion recognition systems have practical
applications across multiple industries:
- Healthcare: Monitoring mental health and
stress levels.
- Marketing: Tailoring advertisements to
consumers' emotional responses.
- Security: Identifying suspicious behavior
through emotional cues.
ETHICAL CONCERNS
Privacy Issues: Emotion recognition systems
require sensitive personal data such as facial
expressions, voice recordings, or physiological
signals. The collection and storage of this data
can raise privacy concerns.
Misuse of Data: The potential for unethical
use, such as surveillance or manipulation (e.g.,
targeted advertising based on emotional
states), creates significant ethical dilemmas.
REAL WORLD
EXAMPLES
For example, by looking at this person's
facial expression. We can comment on
the emotional state. For example, from
his furrowed eyebrows and displeased
look, it can be inferred that he is
currently angry or worried about a
situation. We can also notice his
tension from his body position.
THANK YOU FOR LISTENING!
REFERANCES
https://www.academia.edu/
40049452/%C4%B0nsan_Ses_Renklerinin_Duygudurum_%C3%9Czerine_Etkileri_Doktora_Tezi_The_Effects_of_the_Human_Tone_Colours_On_Mood_
Phd_Thesis_
https://www.drhaldunoguz.com/duygularin-sese-etkisi
https://dergipark.org.tr/tr/download/article-file/2578244
https://evrimagaci.org/duygular-ve-yuz-ifadeleri-yuzumuz-duygularimizi-nasil-ifade-eder-7555?
srsltid=AfmBOooyjq6uXSv71od5F8ySXRapZlqvpxsygRKZR2dUngun6kSQnrXd
Liao, S., & Hamilton, G. (2021). Ethical Challenges of AI in Mental Health Care. AI and Ethics,
2(3), 291-303.
Joubert, M., & McNamara, C. (2020). Artificial Intelligence in Psychology: The Emerging Role
of AI in Clinical and Applied Psychology. Psychology and Cognitive Science, 12(1), 32-45.
https://tr.wikipedia.org/wiki/Paul_Ekman
