 |
 |
 |
| Without Add-on AI module |
|
|
|
V.S.
|
|
|
| Testing condition: Running popular object detection model Yolov5n by ONNX conversion under Ubuntu 22.04 on AIR-150 edge system (CPU: 13th Gen. ULV, Core i5-1345UE, iGPU: Iris® Xe, NPU: Hailo-8 / Kernal: 5.15, Batch size: 1, Data type: INT8) |
|
| |
|
| How-to Guide |
| Choosing the Right AI Acceleration Modules for Your Needs |
|
|
|
 |
|
|
 |
| Security Screening |
|
|
|
 |
| EAI-1200 |
 |
|
| Hailo-8 M.2 AI Module |
|
|
|
|
 |
 |
|
| 13th Gen Intel® Core™ i3/i5 Fanless Inference System |
|
|
|
|
 |
|
Intel® Atom® x6413E
DIN-Rail Fanless Box PC |
|
|
|
|
|
|
| Application Requirements |
| • |
Higher people-per-hour rate for security screening |
| • |
Seamless Integration with current x86 platform |
|
|
| Benefits |
| • |
Bag screening efficiency up to 4,000 people per hour (pph) |
| • |
10 times faster than traditional metal detection |
| • |
Easily build object detection models with Hailo AI toolkits |
|
|
|
|
 |
| Robotics & AMR |
|
 |
| EAI-2100 |
 |
|
Intel® Arc™ A370E
MXM 3.1 Type A GPU Card |
|
|
|
|
|
 |
|
| 13th Gen Intel® 3.5" Single Board Computer & MXM GPU carrier board |
|
|
|
|
 |
|
| 12th/13th/14th Gen Intel® MXM GPU Inference System |
|
|
|
|
|
|
| Application Requirements |
| • |
Lower latency for real-time detection |
| • |
H.264/H.265 media codec support |
| • |
Anti-vibration required for moving around |
|
|
| Benefits |
| • |
Slim MXM GPU & local AE support for seamless integration |
| • |
Anti-vibration under 3 Grms, IEC60068-2-64, random |
| • |
Supports rich media codecs such as H.264/H.265/AV1/VP9 |
|
|
|
|
 |
| Traffic Monitoring |
|
|
|
 |
| EAI-3101 |
 |
|
Intel® Arc™ A380E
PCIe x16 GPU Card |
|
|
|
|
|
 |
|
| 12th/13th/14th Gen Intel® Core™ Desktop Mini-ITX |
|
|
|
|
|
|
| Application Requirements |
| • |
High throughput for multi-stream AI inferencing |
| • |
Stable 24/7 traffic monitoring |
| • |
Image enhancement under adverse weather |
|
|
| Benefits |
| • |
50% latency reduction via 4-lane traffic analysis |
| • |
Saves 33% on construction fees over previous GPU solutions |
| • |
Compatible with customers' OpenVINO-based software |
|
|
|
|
 |
| Defect Inspection |
|
 |
| EAI-3300 |
 |
|
| Hailo-8 PCIe x16 AI Acceleration Card |
|
|
|
|
|
 |
|
Intel® Core™ 14th Gen
Edge AI Server |
|
|
|
|
 |
|
| 13th Gen Intel® Core™ Edge Mini Server System |
|
|
|
|
|
|
| Application Requirements |
| • |
Lower failure rate than traditional AOI |
| • |
Improved inspection speed for production |
| • |
Scalable AI defect detection |
|
|
| Benefits |
| • |
2X improved throughput at up to 15W power |
| • |
40% increase in inspection speed over traditional AOI |
| • |
Supports major inspection models like Yolo, MobileNet-SSD |
|
|
|
|
 |
|
|
 |
