Product & Technology

• How do Kepler Night Nurse and Kepler NurseAssist detect falls in real time and help reduce them?
  Kepler Night Nurse and Kepler Nurse Assist use anonymized computer vision technology and AI-powered sensors that monitor residents 24/7. When a fall is detected, caregivers receive immediate notification, which helps reduce injuries and improve response time.

  
  

  Read more about response times here
• Why is Kepler Vision's fall detection system more reliable than wearables?

  Unlike wearables, residents where Kepler Night Nurse and Kepler NurseAssist are used do not have to wear anything. They don’t have to remember to put a device on or charge it. Our always-active AI sensors monitor continuously, reliably and immediately detecting every fall.
  
  Read our customer reviews here

• How do Kepler Night Nurse and Kepler NurseAssist support fall prevention before incidents happen?

  The system recognizes early risk indicators, such as sitting at the edge of the bed, getting out of bed or getting up in bed. Caregivers are proactively alerted so that preventive measures can be taken.
  
  Read more about Kepler Night Nurse and Kepler NurseAssist here.

• Are Kepler Night Nurse and Kepler NurseAssist really always active, even at night?

  Yes. Daylight is used during the day, and infrared light is used at night to provide 24/7 monitoring. Thus, residents are always safely monitored without disturbing their sleep.

  Kepler NurseAssist always operates on cameras equipped with infrared light.

• What is anonymized digital supervision and how does it protect residents' privacy?

  The software processes video footage locally and converts it into anonymized data where faces are automatically blurred. The analysis focuses only on body movement, not identity. This ensures dignity, privacy and regulatory compliance.

• How do Kepler Night Nurse and Kepler NurseAssist provide real-time notifications for falls or abnormal behavior?

  When the AI detects a fall or unusual activity, notifications are immediately sent via the nurse call system to employees’ mobile devices. This speeds up care and improves safety.

• What is the difference between an AI camera and a regular security camera?

  An AI camera processes video images directly on the device itself so that human activities and risk situations are recognized locally. With an ordinary camera, this processing often does not occur or must occur externally (e.g., in the cloud or on a server).  The advantage of AI cameras is that video footage does not have to leave the camera, offering a significant privacy benefit.

• What are the numbers on the accuracy of the system? How often does it give a false report (false positive) and how often does it miss a report (false negative)?

  For Kepler Night Nurse (server application):

  For man-down alarms, Kepler Vision receives on average one missed man-down per month. Approximately ten false-positive man-down reports reported.

  For Kepler NurseAssist (embedded application):

  The software running on the camera uses the same underlying models. Therefore, the performance statistics should be similar.

• How does the system determine that a notification should be sent? What is the model logic?

  Users can customize any type of notification or alarm. Users select which notifications they want to receive.

• Are there scenarios or environments in which the system performs less effectively? What are the "ideal" conditions to achieve the highest effectiveness?

  The ideal environments are elder care facilities, mental health facilities and hospitals that use a nurse call system (Nurse Call System). It is ideal when the nursing staff has basic knowledge of how a call system works, combined with training on how notifications are generated.

• How are improvement opportunities identified, and what happens when performance declines? Is performance likely to decline over time, or will it remain consistent?

  Performance typically improves during the implementation phase and then increases further when software updates are installed.