Blog post

Announcing Support for Advantech AOM-2721 in Edge Impulse

edge ai
By Edge Impulse Team
Announcing Support for Advantech AOM-2721 in Edge Impulse

Edge Impulse now offers official support for the Advantech AOM-2721 Open Standard Model (OSM). This ultra-compact, high-performance module is now fully integrated into the Edge Impulse ecosystem, making it easier to build, deploy, and run AI at the edge.

What makes the AOM-2721 stand out

At the heart of the AOM-2721 is the Qualcomm Dragonwing™ QCS6490, which delivers robust AI performance through a combination of high-performance Kryo cores — one Gold core up to 2.7 GHz, three Gold cores at 2.4 GHz, and four Silver cores at 1.9 GHz. 

Alongside the CPU, the module features an Andreo VPU 633 capable of 4K30 encoding and decoding, an Andreo GPU 643 with OpenGL ES 3.2 and OpenCL 2.0 support, and a dedicated 12 TOPS NPU for accelerated machine learning inference. This hardware makes the AOM-2721 well-suited for demanding AI workloads at the edge.

The module also offers generous memory and connectivity options. With 8 GB of LPDDR5 RAM operating at 8533 MT/s, it provides ample bandwidth for both AI and multimedia tasks. Developers have access to a wide range of interfaces, including various types of display, USB, PCIe, audio, serial, and camera ports. This combination of performance and I/O flexibility makes it a versatile platform for a wide range of applications.

Advantech EPC-R2860 using the AOM-2721

Getting Started with Edge Impulse

Visit the Edge Impulse docs to walk through the process of preparing the AOM-2721:

From there, you can use the AOM-2721 in Edge Impulse to:

Why use the AOM-2721 with Edge Impulse


The Advantech AOM-2721 OSM paired with Edge Impulse provides a solid, ready-to-use foundation for building edge AI applications in areas such as real-time vision, object detection, voice recognition, and anomaly detection.

Get started with your free developer account to explore what you can do with the AOM-2721 and much more.

Comments

Subscribe

Are you interested in bringing machine learning intelligence to your devices? We're happy to help.

Subscribe to our newsletter