kmitl67056109/test-full-dental-ergonomics-model

Full-Body Dental Ergonomics AI

This model is an unsupervised ensemble designed to assess ergonomic risk in dental professionals. Unlike standard models that only look at one metric, this system contains four separate AI agents, each trained on the specific biomechanics of a different body part.

Capabilities

It processes raw IMU sensor data to detect dangerous posture in:

1. Neck (Cervical Flexion)
2. Upper Back (Trunk Flexion)
3. Right Arm (Shoulder Elevation)
4. Left Arm (Shoulder Elevation)

How it Works

For each body part, the system uses a consensus voting mechanism:

• Voter 1 (RULA Rule): Flags angles exceeding biomechanical thresholds (e.g., >20° for Neck, >45° for Arms).
• Voter 2 (K-Means): Clusters habits into Low, Medium, and High deviation.
• Voter 3 (GMM): Probabilistic assessment of abnormal posture.

How to Use

Download full_body_ergonomics_model.pkl and load it with joblib. The file is a dictionary containing the 4 sub-models.

import joblib
models = joblib.load('full_body_ergonomics_model.pkl')

# Access the Neck model
neck_model = models['head.csv']
neck_kmeans = neck_model['kmeans']