Aiii Book a Demo

AI Foreseeing Parkinson's Disease

In one sentence: Aiii uses computer vision and deep learning (OpenPose motion analysis) to assist early recognition and long-term tracking of Parkinson's disease —— simply record a video in the App to detect tremor and bradykinesia, and digitize MDS-UPDRS scoring and medication-response tracking (supportive reference, not a medical diagnosis).

Parkinson's Time Machine: Foreseeing the Future

Building a tremor-free tomorrow together with Aiii

Protecting brain health, foreseeing a better lifeApril 11 each year is "World Parkinson's Day," commemorating the British physician who first described Parkinson's disease —— Dr. James Parkinson. This day reminds us to pay attention to neurological health, especially the early recognition of and intervention in neurodegenerative diseases.

Parkinson's disease is a common neurodegenerative disorder that mainly affects the central nervous system. Early symptoms include hand tremor, slowed movement, muscle rigidity and balance difficulties. Because the early symptoms are hard to notice, diagnosis is often delayed.

To help with earlier identification and intervention, we have introduced AI Smart Parkinson's Detection technology. With the support of technology, we integrate advanced computer vision and deep-learning algorithms to improve the efficiency and accuracy of disease recognition, giving patients a greater window of opportunity for treatment.


Empowered by AI, we are building a cross-disease health-monitoring ecosystem that ushers the prevention, diagnosis and management of neurodegenerative and chronic diseases into a new era of precision, personalization and proactive intervention. We hope this technology can break the traditional boundaries of disease management, shifting from reacting to illness toward managing health, and bring wellbeing and hope to an aging society worldwide.

aiii_official-website-case_world-parkinsons-day-1

Step 1 Upload a Video

  • Record within the App or select and upload a video from your photo library
  • OpenPose detects in real time whether key body nodes deviate from the preset ranges; if a deviation is detected, the interface instantly prompts the correct way to adjust
  • Checks the uploaded video and notifies the patient whether the upload was successful

Step 2 Integrated Clinical Assessment Platform

  • Digitized UPDRS scoring system: integrates the MDS-UPDRS motor examination into a digital scoring interface
  • Time-series comparison: presents a patient's historical assessment data to support longitudinal comparison
  • Medication-response tracking: records ON/OFF state assessments to optimize medication-adjustment strategies
aiii_official-website-case_world-parkinsons-day-2
aiii_official-website-case_world-parkinsons-day-3

Step 3 Capture the Patient's Body Nodes

  • Use OpenPose to extract key points
    • Reduce data dimensionality: using OpenPose to extract the key points of the human body greatly reduces the amount of data to process, as only the coordinates of the key points in each frame need to be handled rather than the full image of every frame.
    • Focus on motion features: OpenPose is specifically designed to extract human body key points, which is very useful for analyzing features such as body movement and posture. In Parkinson's detection, features such as tremor and bradykinesia are often expressed through limb movement, and using key points allows these movements to be captured more precisely.
    • Enhance the model's generalization ability: through key points, the model can learn motion features that are unaffected by the background or other external factors, which helps improve the model's generalization to different scenarios.
  • Organize features such as the coordinates, velocity, displacement distance and timing of each of the user's body nodes, and apply a Long Short-Term Memory (LSTM) model to predict the severity of the user's Parkinson's symptoms
Pharmaceutical
Pharmaceutical Solutions
Learn More
Enterprise
Enterprise Solutions
Learn More
E-commerce-1
E-commerce Solutions
Learn More

Frequently Asked Questions

How does AI help with early recognition of Parkinson's disease?
Using computer vision and deep learning (OpenPose key-point detection), it analyzes motion features such as limb tremor and bradykinesia to assist earlier identification, giving patients a window of opportunity for treatment (not a medical diagnosis).
How do I use it? Do I need a wearable device?
No. Simply record or upload a video within the App, and the system uses OpenPose to detect in real time whether key body nodes deviate from the preset ranges.
Can it track disease progression and medication response?
Yes. It digitizes MDS-UPDRS motor scoring, performs time-series comparison of historical data, and records medication ON/OFF states to help optimize medication-adjustment strategies.
How is the privacy of personal health footage protected?
The system only captures the coordinates of human body key points rather than the full image, greatly reducing data dimensionality, and processes data on a privacy-first basis.
Will this replace a physician's diagnosis?
No. The AI is a supportive and long-term tracking tool; the final diagnosis and treatment are still determined by qualified medical professionals.

Contact US

contact