Deploying Machine Learning Techniques for Human Emotion Detection.

Area of Science:

• Computer Vision
• Artificial Intelligence
• Human-Computer Interaction

Background:

• Emotion recognition is crucial for human-robot interaction and robotic vision.
• Facial expressions are a primary modality for detecting human emotions.
• Existing methods require real-time, accurate emotion detection for practical applications.

Purpose of the Study:

• To present a real-time approach for emotion detection.
• To implement and deploy this system in robotic vision applications.
• To enhance the accuracy of emotion detection through advanced feature engineering.

Main Methods:

• Utilized MediaPipe face mesh for real-time key point generation.
• Employed mesh generator and angular encoding for key point representation.
• Applied Principal Component Analysis (PCA) for feature decomposition.
• Classified emotions using Support Vector Machine (SVM), k-Nearest Neighbor (KNN), Naïve Bayes (NB), Logistic Regression (LR), Random Forest (RF), and Multilayer Perceptron (MLP).

Main Results:

• Achieved a superior human emotion detection accuracy of 97%.
• Demonstrated effectiveness across various datasets and evaluation metrics.
• Outperformed existing methods in emotion recognition tasks.

Conclusions:

• The proposed real-time emotion detection approach is highly accurate and effective.
• The system is suitable for deployment in robotic vision and interactive communication.
• The combination of MediaPipe, PCA, and ML classifiers provides a robust solution for emotion recognition.