Body-Earth Mover's Distance: A Matching-Based Approach for Sleep Posture Recognition

Area of Science:

• Biomedical Engineering
• Human-Computer Interaction
• Sleep Science

Background:

• Sleep posture significantly impacts sleep quality and pressure ulcer prevention.
• Body pressure analysis is a common technique for recognizing sleep postures.

Purpose of the Study:

• To propose a novel matching-based approach, Body-Earth Mover's Distance (BEMD), for accurate sleep posture recognition.
• To evaluate BEMD's effectiveness by comparing it with existing methods.

Main Methods:

• BEMD treats pressure images as weighted 2D shapes.
• It combines Earth Mover's Distance (EMD) and Euclidean distance for similarity measurement.
• Sleep posture recognition is based on posture similarity, not specific feature detection.

Main Results:

• The pilot study involved 14 participants across six different postures.
• BEMD achieved a high accuracy rate of 91.21% in sleep posture recognition.
• This represents an 8.01% improvement over previous methods.

Conclusions:

• BEMD offers a more effective and accurate method for sleep posture recognition.
• This approach has significant potential for enhancing sleep quality assessment and pressure ulcer prevention strategies.