Computed Tomography (CT)
Magnetic Resonance Imaging (MRI)
Population ageing and the rising demand for early-stage diagnosis are placing increasing demands on medical imaging to provide faster, more efficient and enhanced diagnosis accuracy.
As a result of the rapid development of deep learning algorithms and GPU-accelerated computing, AI is considered to be a catalyst in fulfilling the growing demand of medical imaging. According to medical market research reports, the market for Artificial Intelligence in medical imaging is expected to grow at a CAGR of +30% from 2021 to 2026 .
With over two decades of expertise and experience in optimizing edge computing performance, ADLINK offers GPU solutions to assist medical imaging equipment manufacturers worldwide in achieving consistently clear and precise medical imaging utilizing AI applications.
Challenges of Medical Imaging
Time-consuming Image Reconstruction
Medical imaging involves a series of complex signal and image reconstruction processes that convert raw data detected by X-ray or ultrasound sensors into 2D cross-sectional or 3D volumetric images. This time-consuming image processing and the increasing demand of early-stage diagnosis overwhelms the available bandwidth of medical imaging providers.
Inconsistent Image Quality
Image quality is the key to precise diagnosis. Noisy images can lead to misdiagnosis or result in the need for repeated scanning of patients.
Increasing Demand for Equipment Portability
For patients undergoing critical treatment, transfer to standard CT scanning room would expose the patients to potential risks and dangers. Therefore, hospitals increasingly rely on portable medical imaging devices.
Industrial-Grade GPU-Enabled Platforms for Medical Imaging
Embedded MXM GPU modules powered by NVIDIA Turing™ architecture
One of the smallest industrial-grade GPU-enabled platforms in the world
Lifespan up to 5-years
Benefits of Industrial-Grade GPU-Enabled Platforms
Direct memory access speeds up data transmission with reduced latency
ADLINK’s industrial-grade GPU-enabled platforms offer high system responsiveness with embedded GPUs. In contrast with graphics cards for gaming, embedded GPUs feature NVIDIA GPUDirect® remote direct memory access (RDMA) which can deliver an 80% boost in data throughput by giving external data sources direct access to the GPU’s external memory. With this feature, data will not be copied into a CPU’s memory before reaching the GPU, significantly decreasing data transmission delay and system latency.
Accelerate medical image reconstruction with AI assistance to ensure consistent quality
ADLINK’s embedded MXM GPU modules based on NVIDIA Turing™ architecture offer computing acceleration for medical imaging. The Turing architecture enhanced CUDA® Cores which can deliver a 50% improvement in performance compared with the previous Pascal™ generation GPUs. Ray tracing, which is fundamental to 3D volumetric imaging, can now be rendered in real time on a single GPU with hardware accelerators known as RT Cores. The integrated Tensor Cores provide acceleration for matrix computations at the heart of deep learning neural network training and inferencing operations, which would benefit artifacts removal, contrast adjustment, and sharpness enhancement for clearer medical images.
Compact yet powerful to fit in space-limited portable medical imaging equipment
ADLINK’s DLAP-3000-CF series is one of the smallest industrial-grade GPU-enabled platforms in the world. It’s only 3.2 liters of volume, making it a perfect fit for space-limited portable medical imaging equipment.
Considering the substantial computing workloads image reconstruction requires and the massive amount of heat this generates, the high performance-per-watt and extended temperature options of ADLINK’s embedded MXM GPU modules allows the DLAP-3000-CF series to dictate the speed and accuracy of image reconstruction. Additionally, it's quiet active cooling design keeps the system stable with noise levels below 40db that no one would notice in the diagnosis
5-years lifespan to ensure long-term service for your investment
As an NVIDIA® Quadro® Embedded Partner, ADLINK offers embedded graphics cards with a 5-years product life. This 3-times the lifespan compared to the 1.5-years continuity of non-embedded graphics cards. Given the stringent, time-consuming development, verification and validation processes of medical devices, availability of key components like GPUs could mean the difference between continuity and end of life for medical imaging equipment.
ADLINK’s Industrial-Grade GPU-Enabled Platforms
|Processor||8th/9th Ge Intel® Core™i7/i5/i3 Processors||9th Gen Intel® Xeon®,
|Chipset||Intel® H310 Chipset||Intel® Q370 Chipset||Intel® H310 Chipset||Intel® C246 Chipset|
|Memory||DDR4 2666/2400MHz, 2x SO-DIMM, up to 64GB (dependent on CPU)||DDR4 2666/2400MHz, 2x SO-DIMM, up to 32GB||DDR4 2400MHz, dual SODIMMs, up to 64GB with ECC option|
|GPU||Embedded MXM GPU Module Type A/B/B+||Full-Height, Full-Length PEG card|
|Display||6x DP (2 from CPU, 4 from MXM)||1x DVI-D, 1x DisplayPort, 1x HDMI||2x DP++, 1x DVI-I|
4x USB 3.1,
4x USB 2.0.
6x USB 3.1,
2x USB 2.0.
4x USB 3.1,
2x USB 2.0.
3x USB 3.1,
3x USB 2.0,
1x internal USB 2.0 dongle
|Expansion||-||-||2x PCIe x4 expansion slot||1x PCIe x16 slot||
Physical: 2x PCIe x16, 2x PCIe x8, 1x PCIe x4;
Signal: 2x PCIe x8, 2x PCIe x4, 1x PCIe x1
2x 2.5" SATA 6Gb/s drive bays,
1x SATA 6Gb/s via M.2 B key
2x 2.5" SATA 6Gb/s drive bays,
2x SATA 6Gb/s via M.2 M & B key connector,
Intel® RST RAID Support
|2x 2.5" SATA 6Gb/s drive bays||
4x 2.5" SATA 6Gb/s hot swappable drive bays,
Supports RAID 0, 1, 5, 10
Optional: 240W (12V/20A) AC/DC adapter
100 to 240 VAC,
300W / 500W Built In PSU
Optional AC input 1000W
(H x W x D)
|235 x 182 x 75mm||235 x 182 x 75mm||235 x 182 x 130mm||220 x 300 x 150mm||210 x 210 x 350mm|
ADLINK’s Medical Imaging Projects & Customer Benefits
Mobile X-ray, C-arm
Enhanced image quality with GPU computing acceleration for medical imaging.
Maximum system uptime with low power consumption.
Increased mobility with size-, weight- and power-optimized design.
Meet the ever-growing demand for high-resolution images with reduced I/O latency and enhanced system responsiveness.
High and fast response for diagnostic and clinical use.
Integrated AI to help detect anomalies and enable faster diagnosis.
Faster 3D image reconstruction for enhanced patient experience.
Improved image reconstruction with AI to enhance treatment and patient outcome.
Partnership between NVIDIA and ADLINK
ADLINK is a premier embedded computing company with access to NVIDIA’s embedded GPU roadmap.
As NVIDIA Jetson Elite and OEM Preferred Partner, ADLINK gets direct support and early access from NVIDIA.
We are ready to accelerate your success ─ private labeling and semi-customization are welcome.