Drug diversion in hospitals is a serious concern, facilitated in part by inadequate security controls that allow unauthorized staff to access the pharmacy stock. To help thwart this problem and improve medication management, Taipei, Taiwan-based Cyberlink Corp. and Imedtac Co. Ltd. are teaming up to bring facial recognition technology to the hospital medicine cabinet.
Specifically, Cyberlink is integrating its artificial intelligence (AI)-based Faceme technology into iMedtac’s Automated Dispensing Cabinet to create a smart medicine cabinet powered by internet of things (IoT) and facial recognition capabilities. The medicine cabinet is deployed through an AIoT iHospital service platform to increase overall pharmaceutical management within the hospital.
The collaboration with Imedtac is Cyberlink’s first foray into the health care field. “These products should be first deployed in hospitals in Taiwan and neighboring Asian markets, with potential for global distribution,” Richard Carriere, senior vice president and general manager at Cyberlink, told BioWorld.
He added that a number of other partnerships, business deals and proofs of concept are in progress covering a range of health care-related use cases around authentication, access control, security and visitor and patient identification.
“The rapid growth of AI technologies enables innovation across multiple facets of health care, driving breakthroughs in treatment, safety and quality of patient care,” said Jau Huang, founder and CEO of Cyberlink. “The integration of Faceme AI facial recognition into Imedtac’s smart medicine cabinet is a good example of how AIoT solutions can drive transformation of the medical industry and deliver safer, more efficient and better care.”
Powered by AI and deep-learning algorithms, Faceme offers high-precision, real-time facial recognition, with a 99.82% accuracy rate over the Labeled Faces in the Wild, a database of photographs considered a public benchmark for facial verification. In the latest face recognition vendor test (FRVT), conducted by the U.S. National Institute of Standards and Technology, Cyberlink ranked 12th among all participants in FRVT 1:1 (WILD 1E-4), Carriere said.
Moreover, Faceme is built to run on Windows, Linus, Android and iOs operating systems with a variety of edge devices – those that serve as point of entry into an enterprise – and hardware configurations. “The ability to run Faceme on edge devices ensures very precise facial recognition within a few milliseconds, and delivered in a highly secure manner, as there is no need for a cloud connection or heavy server-based processing,” Carriere said.
Loss of drugs and deliberate thefts in hospitals and health systems are on the rise, posing a risk to patient safety as well as lawsuits and regulatory repercussions. In August 2018, the University of Michigan Health System agreed to pay the U.S. government $4.3 million to settle allegations that it mishandled opioids, leading to the theft of roughly 16,000 hydrocodone pills between May 2011 and January 2012.
Earlier this year, the National Association of Drug Diversion Investigators reported the theft of more than 18,000 pills, mostly opioids, from Beverly Hospital in Boston, a part of Lahey Health. The culprit was a pharmacy technician who changed the computer records to indicate that the drugs were outdated. Such incidents are enabled, in part, by control procedures that rely on keys and passcodes to access medicine supplies.
By implementing Cyberlink’s Faceme AI facial recognition engine, Imedtac’s Automatic Dispensing Cabinet brings AI and IoT technology into medical care, enhancing the safety of prescription drug delivery,” said Ken Yu, Imedtac’s founder.
Depending on which report you read, the global smart hospital market is expected to reach anywhere from $58.8 million to $106.5 billion by 2025, fueled by growth in connected medical devices, adoption of IoT technologies and cloud computing services, and hospitals’ quest for more efficient, cost-effective workflow solutions. In addition to medication management, key use cases include electronic health records and clinical workflow, medical connected imaging, outpatient monitoring and medical assistance.
Carriere said it’s too early to predict facial recognition’s market potential in health care, but noted that it is a vertical market with several high value-added applications for the technology. “We are quite active already, and we have no doubt it will be substantial,” he said.