Advantech and Leopard Imaging Advance Machine Vision and Physical AI with Integrated Solutions at AUSA 2026

Taipei, October, 2026 –— Advantech (TWSE: 2395) is collaborating with Leopard Imaging® to advance machine vision and Physical AI applications by combining robust edge computing platforms with a broad camera portfolio and advanced imaging expertise. This synergy significantly simplifies the complexity of embedded camera integration, delivering a seamless out-of-the-box experience for vision-related project development.   

Debuting at AUSA 2026, the partnership’s first major milestone features an integrated demonstration combining Advantech’s MIO-5381 3.5” single-board computer, powered by Intel® Core™ Ultra Series 3 processors, with Leopard Imaging’s AR0234 GMSL2 cameras (LI-AR0234CS-YUV-GMSL2-120H). The live demonstration brings multi-camera capture and AI inference workloads together on a single, high-performance platform, showcasing the potential to deliver superior imaging capabilities within space-constrained footprints. This collaboration marks the first milestone in the partnership, with plans for deeper cooperation on cross-platform embedded camera support and an expanded product portfolio. The initiative is designed to address the growing demand for diverse vision-based applications across industries, including healthcare, factory automation, smart retail, railway transportation, and smart cities in the Physical AI era.

Collaborative Vision Integration at AUSA 2026

At AUSA 2026, the joint live demonstration connects four Leopard Imaging AR0234 GMSL cameras to the MIO-5381, with each camera delivering live video streams at 1920 × 1200 resolution and 30 frames per second. Alongside these camera feeds, eight separate pre-recorded surround-view video streams run person detection using OpenVINO™ and the platform’s integrated GPU, with detection results displayed on screen.   These concurrent workloads demonstrate the platform’s ability to support low-latency image capture, hardware-accelerated video decoding, and GPU-based AI inference. All image processing and inference in the demonstration take place locally, without sending video data to the cloud.

Joint Image Excellence & Application-Tailored Engineering

Leopard Imaging operates an image quality testing laboratory in Fremont, California, using DXOMARK evaluation methodologies and offering image assessment, calibration, and tuning services. Combined with Advantech’s platform integration capabilities, this collaboration supports project-specific evaluation of image quality under different lighting and capture conditions.   For outdoor, mobile, and other demanding environments, Leopard Imaging can also evaluate cameras with water- and dust-resistant designs while considering temperature, vibration, and installation requirements at the system level. This approach incorporates application and deployment requirements into solution development, helping ensure that camera solutions meet the demands of real-world operating environments.

U.S.-Based Support and Continued Ecosystem Collaboration 

Leopard Imaging’s U.S.-based engineering and manufacturing capabilities support projects in the U.S. market. For projects with specific procurement or sourcing requirements, manufacturing arrangements, country-of-origin documentation, and other related considerations can be evaluated based on the selected products and supply configurations. Advantech will continue to collaborate with Leopard Imaging and additional silicon partners to expand camera integration across different computing architectures, providing more development options for machine vision and Physical AI applications.

Visit Advantech at AUSA 2026, Booth #2266, to explore the integrated demonstration and learn more about the collaboration between Advantech and Leopard Imaging.