Video analysis helps caregivers monitor and react to critical events
Video-pattern-recognition programs enable automated patient monitoring for diagnostic and safety tasks.
With the average lifespan of citizens continuing to rise, so will the number of patients in hospitals and long-term-care facilities. Automated caregiving can help ease the burden of resident monitoring on healthcare personnel. Researchers at the Smart Media and Intelligent Living Excellence (SMILE) Laboratory (National Cheng Kung University, Tainan City, Taiwan) are applyingautomated video and physiological-signal analysis for healthcare applications, in particular systems for respiration and human-behavior analysis with critical event detection.
The team of Pau-Choo Chung, Yung-Ming Kuo, Chin-De Liu, and Chun-Rong Huang usedhidden Markov model (HMM) techniques to monitor critical events in a respiration application as well as wheelchair detection by analyzing respiration motions and human behavior models, respectively.
Abnormal respiration patterns were detected by a series of subsystems and, usingnear-infrared (NIR) imaging, analyzed to measure the periodic rising and falling of the subject’s chest or abdomen. This enabled the system to automatically alert caregivers for diagnosis or intervention if respiration changed to a critical level.
Real-time video monitoring also used HMM techniques to detect abnormal, dangerous events by establishing a base, normal pattern of wheelchair user behaviors for comparison during surveillance in a healthcare facility.
The team plans to advance its research to develop application-specific services that can be automatically triggered when disabled people are detected in wheelchairs, for example, when a patient attempts to leave a chair to get into bed without assistance, which could help prevent patient falls and improve quality of care.
For the full story, visit SPIE’snewsroom.
-- Posted byVision Systems Design