Full Cloud Support Full Cloud Support

  • Fully integrated with a medical grade HIPAA/GDPR compliant cloud backend
  • Provides capabilities to upload, annotate, share and manage surgical videos
  • Implements full life cycle AI/ML Ops for continuous learning

Edge AI Edge AI

  • Delivers real-time inferences with glass-to-glass latencies as low as 150ms using models optimized to run on the QCS8250 SoC heterogenous compute
  • Enables interoperative use of machine/deep learning models for image segmentation, recognition, and other innovative applications

Robust video processing Robust video processing

  • Implements a wide range of multi-camera or multi-display medical video algorithms for encoding, decoding, image processing, and more
  • Accelerates the algorithms leveraging the GPU, multi-vector-DSPs and other embedded hardware accelerators
  • Runs on Android OS with rich multimedia middleware frameworks and analytics engines like FastCV

Designed for medical environment Designed for medical environment

  • Compliant with MOPP and IEC 60601-1 standard for safety and essential performance of medical electrical equipment
  • Follows FDA requirements and IEC62000 immunity standards for EMI, EMC, and ESD
  • Designed as per FDA guidelines of hardware-based uninterrupted video link between camera and display through watchdog in case of any software malfunction

Multi-layer security strategy Multi-layer security strategy

  • Hardware – Provides encryption through TrustZone(TZ) feature of QCS8250 by isolating critical processes that run on the SOC from other kernel processes
  • Cloud– Enables role based access control policy and device provisioning mechanism to enable only authorized entities to communicate with the cloud
  • Chipset– Offers end to end encryption with an external crypto chip that is fully isolated from the SoC and acts as a secure store for all security certificates

Comprehensive development platform Comprehensive development platform

  • Includes customizable medical-grade carrier board, powerful QCS8250 SOM, SDKs, and Board Support Package (BSP) with optimized bootloader and drivers
  • Leverages ARI device management plugins for connecting to the cloud and SNPE SDK for implementing trained ML models on the platform
  • Access to full set of design collaterals like system specifications, user guides, and carrier board CAD files

High-performance interfaces High-performance interfaces

  • Industry-standard custom SMARC and M.2 B key connectors for hardware extensions such as AI accelerators, compatible storage (NVMe memory), and WWAN
  • Provides multiple types of interfaces – DVI, USB, HDMI, PCIe and MIPI-CSI to support video input from any type of camera and send output to any type of display

Customizable platform Customizable platform

  • Allows customization of the Splash screen and Android launcher of platform to reflect the branding guidelines of the OEM
  • Ability to port OpenCV and other implementation of image processing algorithms to work on onboard GPU and DSP
  • Allows customization of cloud connectivity device agents to connect with public cloud IoT platforms such as Azure/AWS or even with OEM specific private cloud