By 2030, it is projected that one in every six people on Earth will be aged 60 or older . Meanwhile, life expectancies are predicted to continue to rise worldwide. These converging trends are expected to create an unprecedented strain on healthcare facilities, possibly more than they are currently equipped to handle.
Fortunately, many technologies offer a clear path forward. One such solution is AI-based visual patient monitoring, where edge systems track patient movement and vital signs. This allows medical personnel to identify risks before they become a concern.
At NXP, we equipped engineers with the i.MX 95 applications processor and Ara240 discrete Neural Processing Unit (DNPU) to enable next-generation clinical workflow solutions.With these tools, the industry can meet imminent healthcare demands and deliver better patient outcomes ultimately saving lives.
Today’s Clinical Challenges
The healthcare industry is already under significant strain. Research shows that since 2022, over 138,000 nurses have left the U.S workforce . For those who remain, a common goal is to reduce daily stress and improve unsustainable workloads so that every patient receives attentive care.
One of the most impactful ways to support clinical teams is by addressing alarm fatigue. Approximately 85% of nurses report feeling overwhelmed by the sheer volume of medical device alerts , many of which not actionable. When a system generates too many false alerts, it can lead to desensitization and divert workers away from more immediate needs.
Ultimately, the industry is suffering from overreliance on labor-intensive monitoring processes that are difficult to scale. Rather than scaling staff to meet the imminent influx of patients, healthcare organizations are tasked with improving the efficiency of existing staff increasing their burden.
Traditional systems rely on the patient to alert personnel to issues they are experiencing using a nurse call button. Unfortunately, n some cases, the patient may be incapable of initiating the alert. Methods for informing hospital personnel of changes in patient status include alarms from monitoring equipment—BP, IV, heart monitors, etc.—or scheduled staff rounds. Unfortunately, in many cases, the alarms are not heard or responded to. Video patient monitoring removes the disorganization or delayed response to status changes as it provides continuous updates—and in many instances, in real time.
Understanding Vision-Based Patient Monitoring
Traditional monitoring may involve uncomfortable wires and sensors attached to the patient, vision-based systems use sensors and wall - or bed-mounted cameras to monitor patients. However, vision-based patient monitoring offers clinicians a means of non-intrusive, continuous patient observation. Local processing then runs video streams through vision AI algorithms to perform predictive analytics based on patient behavior.
Specifically, such systems can track vital signs, detect anomalous behavior (e.g., falls or visible patient discomfort) and promptly alert healthcare staff to intervene if needed. Notably, since these systems run autonomously, they can deliver truly continuous and reliable monitoring, ensuring that patients get the constant attention and care they need.
Opportunities for Clinical Automation
With contactless patient monitoring, clinicians can automate their workflows to address many of their existing problems. For example, these systems can reduce the number of false alarms generated by traditional monitoring systems. By accurately distinguishing routine patient movements from true emergencies, vision-based systems ensure that nurses receive alerts only when patient assistance is necessary. This gives teams more bandwidth to care for patients while reducing alert fatigue,allowing organizations to utilize human resources more effectivly and managage higher patient volume without requiring more staff.
Similarly, vision-based monitoring unlocks a more connected care environment. Because doctors and specialists cannot always be physically present in every hospital wing, vision AI solutions keep them informed of patient status through notifications sent directly to their devices such as laptops or phones. This allows doctors to perform real-time video consultations and behavioral analysis from anywhere, so thatexpert guidance is always accessible to patients on demand—even when the doctor is off-site
New Workflows Empowered by i.MX 95 and Ara240
Medical-grade monitoring systems need compute solutions that are high-performance enough to execute vision-AI at the edge, while robust enough to maintain the highest standards of safety and reliability.
That’s why we designed the i.MX 95 applications processor to combine high-performance compute with specialized imaging capabilities. For real-time and safety applications , the processor includes an Arm® Cortex®‑M7-M5 application cluster, an Arm Cortex-M7-based real-time core and an Arm Cortex-M33 safety domain. For more advanced applications , an integrated Image Signal Processor (ISP), an Arm Mali Graphics Processing Unit (GPU) and an eIQ Neutron Neural Processing Unit (NPU)collectively provide the system with a powerful base of specialized compute resources.
i.MX 95 applications processor.
For a better experience, download the
block diagram.
For more intensive vision processing tasks, the i.MX 95 pairs perfectly with NXP’s Ara240 DNPU via a high-bandwidth Peripheral Component Interconnect express (PCIe) Gen 4 x4 or USB interface. It delivers up to 30 tera operations per second (TOPS) for 8-bit Integer (INT8) operations and 15 tera floating-point operations per second (TFLOPS) for BF16, all while maintaining a highly efficient 3 TOPS/W profile. To support large-scale AI models for vision workflows, it accommodates up to 16GB of low-power double data rate (LPDDR) 4x/5 memory. With these devices processing video feeds locally, engineers can build patient-monitoring systems that are safe, secure and respond to emergencies in real time.
Saving Lives with Smarter Workflows
As the healthcare industry braces for an aging society, edge AI systems will redefine the standard of care. With NXP’s industry-leading hardware bringing intelligence to patient observation, we can create a more accurate, efficient and responsive healthcare system.
NXP is committed to helping medical OEMs build smarter and more human-centered healthcare environments. Explore our i.MX 95 applications processors and the Ara240 DNPU to inspire your next medical design.
