MBX Debuts New Embedded Edge AI Platforms for Medtech Developers
February 6, 2023
Next-Gen Performance Expands AI Capabilities and Lowers Total Cost of Ownership
ANAHEIM, CA (February 6, 2023) – MBX Systems, a specialized hardware designer and integrator, has expanded its suite of embedded and mobile edge AI platforms for medical technology developers to provide powerful new inferencing performance using next-generation NVIDIA and Intel CPUs. The new hardware building blocks enable ISVs to deliver complex medical devices that support inferencing of multiple camera streams at high frame rates, lower the cost of AI processing at the edge, monitor cameras in multiple hospital rooms, and more.
MBX platforms being showcased at Booth #3635 at the MD&M West trade show that opens tomorrow include:
- The new MBX Optio M100 edge AI embedded device featuring NVIDIA® Jetson™ AGX Orin, providing next-level AI performance as well as lower TCO. The new hardware platform – the first in a larger AGX Orin-based embedded device line – utilizes Ampere’s new ARM-based server architecture and delivers 6 times the AI performance, 1.7 times the CPU performance and 1.4 times the memory bandwidth compared to the NVIDIA Jetson Xavier AGX platform. It can be used in the MBX Kori cart for applications such as patient monitoring, diagnostic imaging and telehealth, and in other complex edge AI systems.
- The new ultra-small embedded or standalone MBX Varion® P2 reference platform that can support multiple cameras for on-device AI and inferencing. Built for big performance in a small footprint, the Varion P2 utilizes an x86-based motherboard and the latest Intel® Core™ processors and contains 1x PCIe x16 slot that can accommodate a single wide graphics card such as NVIDIA RTX and Ampere series, capture cards, network adapters, and more. It supports both Linux and Windows operating systems, features a 10-year lifecycle, and includes an optional VESA mount for installation in large lab and diagnostic equipment. It is also an ideal compute option for the MBX Kori cart and other edge AI solutions.
- A dual-camera inference demo using the NVIDIA Jetson AGX Orin platform on MBX’s Kori™ modular mobile AI cart, enabling models to run simultaneously at 100 FPS by leveraging AGX Orin’s increased performance and faster memory. This allows Kori to support applications requiring concurrent inferencing of camera streams such as conversational AI and fall detection in a hospital room or object detection and body positioning in a surgical environment. Kori can be configured with different workstations, cameras and peripherals such as keyboards, monitors and articulating arms, eliminating the need for ISVs to spend more than a year and $100,000+ for custom hardware design, engineering and plastic mold injection.
Platforms can support AI/ML software building blocks such as NVIDIA Clara Guardian, NVIDIA Clara Holoscan, Google TensorFlow and Intel OpenVINO that provide libraries of tools such as pre-built training models that can slash development time for medical AI, machine learning and computer vision applications by a factor of as much as 10.
Later this year, MBX will be expanding the Optio product line by introducing industrial-grade edge AI computing solutions based on NVIDIA IGX Orin targeting medtech developers building applications for uses such as surgical robotics, genomics, and clinical AI devices enabling cross-referencing of patient data with imaging results in real time.
More information on MBX’s platforms and services for medical applications is available at medical.mbx.com.
7 Reasons to Keep Applications & Data On-premise
ARM-Based Processors Are The Future: Delivering High Performance, Low Cost, Sustainable Computing Solutions
MBX Expands Edge AI Hardware Portfolio with NVIDIA IGX Orin-Based Platforms
AHEAD Engineered Solutions
Why Pre-Trained Models Matter for Machine Learning