AI uses night-vision camera to diagnose sleep apnoea from home

Researchers from Seoul National University, Seoul National University College of Medicine, and Columbia University have developed an AI-driven camera system that can diagnose obstructive sleep apnoea (OSA) from home. The system, called SlAction, uses infrared videos to monitor sleep patterns and has demonstrated an 88% accuracy rate in identifying OSA. This offers an alternative to the traditional method of diagnosis, polysomnography (PSG), which requires overnight hospital admission and the attachment of multiple sensors to the patient. The SlAction system captures minor human movements during sleep that are indicative of OSA.

Review: AI Uses Night-Vision Camera to Diagnose Sleep Apnoea from Home

Obstructive sleep apnoea (OSA), a condition characterized by interrupted breathing during sleep, may soon be diagnosed using an AI-driven model that can be utilized from the comfort of one’s residence. Researchers from Seoul National University, Seoul National University College of Medicine, and Columbia University have developed a camera system that can diagnose OSA from sleep recordings.

Traditionally, OSA diagnosis requires overnight hospital admission and the attachment of up to 20 sensors to the patient for polysomnography (PSG). However, PSG is slow, cumbersome, and has its limitations. To address these issues, the researchers introduced a system that uses an infrared camera to observe individuals suspected of OSA during sleep. Their AI tool, SlAction, can then diagnose the disorder from the video.

The researchers trained the AI using video footage of individuals collected from three hospitals. These videos, marked with professional diagnoses, allowed the AI to learn to identify visual signs of OSA, such as frequent awakenings or gasping. In evaluations, the system achieved an 88% accuracy rate in identifying OSA.

Study Details

Introduction of SlAction: The researchers developed SlAction, an innovative system designed to detect OSA using infrared videos to non-intrusively monitor sleep patterns.

Limitations of Polysomnography (PSG): PSG, the current method for diagnosing OSA, requires an overnight stay at a specialized hospital with multiple sensors attached to the patient. This method is prone to inaccuracies due to the “first-night effect” and discomfort caused by the sensors.

Core Research Focus: The researchers aimed to determine if respiratory events associated with OSA are reflected in human movements during sleep. They analyzed a large sleep video dataset and found correlations between events indicative of OSA and minor human movements during sleep.

Technical Approach: SlAction uses a low frame rate and a large window size for capturing videos, ensuring the system captures slow, long-term motions related to OSA. Privacy is maintained through local processing of all video streams.

Results: Preliminary tests of the SlAction system showed an F1 score of 87.6% in detecting OSA across different environments.

AI-supported OSA diagnosis is just one of the many medical technological advancements seen this year, ranging from restoring speech and movement to stroke and accident victims to developing new drugs and diagnosing Parkinson’s disease from eye images.

Source: DailyAI

Action Items:

1. Conduct further research on the AI-driven model for diagnosing obstructive sleep apnoea (OSA) using infrared videos. Determine its potential benefits and limitations compared to traditional methods such as polysomnography (PSG).

2. Explore the feasibility of implementing the SlAction system in our medical facility or recommending it to relevant healthcare providers. Assess the required infrastructure, equipment, and training for using the system effectively.

3. Identify potential collaborations or partnerships with Seoul National University, Seoul National University College of Medicine, and Columbia University to further understand the development and implementation of the AI tool for OSA diagnosis. Seek opportunities for knowledge exchange and joint research efforts.

4. Evaluate the privacy and security measures implemented in the SlAction system to ensure compliance with relevant regulations and protect patient data. Consult with our IT and legal departments to assess any potential risks and mitigation strategies.

5. Consider the potential cost savings and patient benefits that could result from adopting the AI-driven model for OSA diagnosis. Assess the impact on hospital admissions, waiting times, patient comfort, and overall diagnostic accuracy.

6. Communicate the findings and potential benefits of the AI tool to relevant healthcare professionals, including sleep specialists, pulmonologists, and primary care physicians. Organize presentations or training sessions to increase awareness and understanding of the new diagnostic approach.

7. Monitor further developments in medical technological advancements, particularly in the field of AI-supported diagnostics. Stay informed about other emerging technologies and their potential applications in various medical specialties.


1. Hyung-Sin Kim and his team: Further refine the AI tool, SlAction, by analyzing more video footage and conducting additional testing to improve its accuracy in diagnosing OSA. Seek opportunities for collaboration with other research institutions to expand the dataset and enhance the system’s capabilities.

2. IT department: Assess the technical requirements and feasibility of implementing the SlAction system in our medical facility. Explore any necessary adaptations or integrations with existing infrastructure and ensure the system’s compliance with privacy and security regulations.

3. Legal department: Review the legal and ethical implications of using the SlAction system for OSA diagnosis. Ensure compliance with relevant data protection regulations and consider any necessary consent procedures for patients participating in the video recording process.

4. Research and development team: Investigate potential applications of AI technology in other areas of healthcare, such as stroke rehabilitation, neurodegenerative diseases, and drug development. Stay up to date with the latest advancements and identify opportunities for collaboration or investment in these areas.

Note: These action items and assignments are suggestions based on the meeting notes provided. Please review and modify them as necessary to align with your organization’s goals and priorities.

List of Useful Links:

AI Products for Business or Try Custom Development

AI Sales Bot

Welcome AI Sales Bot, your 24/7 teammate! Engaging customers in natural language across all channels and learning from your materials, it’s a step towards efficient, enriched customer interactions and sales

AI Document Assistant

Unlock insights and drive decisions with our AI Insights Suite. Indexing your documents and data, it provides smart, AI-driven decision support, enhancing your productivity and decision-making.

AI Customer Support

Upgrade your support with our AI Assistant, reducing response times and personalizing interactions by analyzing documents and past engagements. Boost your team and customer satisfaction

AI Scrum Bot

Enhance agile management with our AI Scrum Bot, it helps to organize retrospectives. It answers queries and boosts collaboration and efficiency in your scrum processes.