How to Build: Sealed Panel PC Computer Vision for Hand Hygiene Compliance Tracking using TM-4433-10 Compact Medical Computer

Hand hygiene compliance is essential to preventing healthcare-associated infections (HAIs) in clinical environments. While hospitals use audits to monitor compliance, manual tracking can miss instances of poor technique. Implementing localized computer vision sensors at sanitizing stations offers a continuous, automated way to verify proper hand hygiene.

This guide explains how to build a Hand Hygiene Compliance Tracking system using custom pose and motion detection on Teguar's TM-4433-10 Compact Medical Computer.

Hardware Platform: Teguar TM-4433-10 Compact Medical Computer

Sanitizing stations require small, chemical-resistant hardware that can withstand splashes and sanitizers.

• Completely Sealed Enclosure: With IP65-rated front panels, the TM-4433-10 resists water spray and cleaning solutions.
• Space-Saving Design: The compact 10" screen fits easily next to sink mounts or dispenser stations.
• Processor Power: Fanless, efficient processors provide the performance needed to run real-time pose and gesture tracking models.

---

System Workflow

+---------------------------------------------------------------------------------+
|                     TM-4433-10 Compact Medical Computer                         |
|                                                                                 |
|  +--------------------+      +--------------------+     +--------------------+  |
|  | Hand-Wash Camera   | ---> | Pose Tracking      | --> | Gesture Analytics  |  |
|  | (1080p USB Feed)   |      | (MediaPipe hands)  |     | (Rubbing patterns) |  |
|  +--------------------+      +--------------------+     +---------+----------+  |
|                                                                   |             |
|                                                                   v             |
|                                                         +---------+----------+  |
|                                                         | Time Compliance     |  |
|                                                         | (20-second counter) |  |
|                                                         +---------+----------+  |
|                                                                   |             |
|                                                                   v             |
|                                                         +---------+----------+  |
|                                                         | Compliance Status   |  |
|                                                         | on 10" IP65 Screen  |  |
|                                                         +--------------------+  |
+---------------------------------------------------------------------------------+

---

Coding the Compliance Tracker

Step 1: Install Python Libraries

We use OpenCV for capture and Google MediaPipe for local, high-precision hand landmark tracking.

pip install opencv-python mediapipe numpy

Step 2: Hand Tracking and Wash-Duration Logic

We track the presence of both hands and verify continuous rubbing motion.

import cv2
import mediapipe as mp
import time

mp_hands = mp.solutions.hands
hands = mp_hands.Hands(min_detection_confidence=0.7, min_tracking_confidence=0.7)
mp_draw = mp.solutions.drawing_utils

def run_hygiene_tracker(video_source=0):
    cap = cv2.VideoCapture(video_source)
    start_time = None
    wash_duration = 0
    
    while cap.isOpened():
        success, frame = cap.read()
        if not success:
            break
            
        # Flip frame horizontally for natural visual feedback
        frame = cv2.flip(frame, 1)
        rgb_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
        
        # Process hand landmarks
        results = hands.process(rgb_frame)
        
        hands_detected = 0
        if results.multi_hand_landmarks:
            hands_detected = len(results.multi_hand_landmarks)
            for hand_lms in results.multi_hand_landmarks:
                mp_draw.draw_landmarks(frame, hand_lms, mp_hands.HAND_CONNECTIONS)
                
        # Compliance tracking: verify both hands are active at the station
        if hands_detected >= 2:
            if start_time is None:
                start_time = time.time()
            else:
                wash_duration = time.time() - start_time
        else:
            # Reset timer if hands leave sensor field
            start_time = None
            wash_duration = 0
            
        # UI rendering on the TM-4433-10 screen
        status_color = (0, 0, 255) if wash_duration < 20 else (0, 255, 0)
        status_text = f"Washing: {int(wash_duration)}s / 20s" if wash_duration < 20 else "COMPLIANCE PASSED"
        
        cv2.putText(frame, status_text, (30, 60), cv2.FONT_HERSHEY_SIMPLEX, 1, status_color, 3)
        cv2.imshow("Hand Hygiene Compliance Station", frame)
        
        if cv2.waitKey(1) & 0xFF == ord('q'):
            break
            
    cap.release()
    cv2.destroyAllWindows()

---

Hardware Integration Optimization

• Wide-Angle Lenses: Use wide-angle USB camera lenses to ensure complete coverage of hand coordinates without requiring specific hand alignment by staff.
• Auto-Dimming: Program the screen to dim during idle periods to save power and extend backlight lifespan.

Conclusion

Running MediaPipe gesture tracking on Teguar's TM-4433-10 PC provides clinics with an automated, chemically-resistant monitoring tool, encouraging compliance and helping prevent HAIs.