Fathom Neural Compute Stick Enables Mobile Devices to React Cognitively

Print Friendly, PDF & Email

Intel-owned Movidius has introduced a fascinating new device called the Fathom Neural Compute Stick, a modular deep learning accelerator in the form of a standard USB stick.

The Fathom Neural Compute Stick is the first of its kind: A powerful, yet surprisingly efficient Deep Learning processor embedded into a standard USB stick. The Fathom Neural Compute Stick acts as a discrete neural compute accelerator, allowing devices with a USB port run neural networks at high speed, while sipping under a single Watt of power. The Fathom Neural Compute stick can also be used in prototyping and validating neural networks, moving them from the PC environment to actual end products like smart cameras, drones and mixed reality headsets.

Featuring a full-fledged Myriad 2 VPU, the Fathom Neural Compute Stick not only enables rapid prototyping, but also delivers high levels of neural network compute to existing devices via a USB port. Thanks to the ultra-low power Myriad 2 VPU, the Fathom Neural Compute Stick does not require an external power supply, and can accelerate neural network performance for a wide range of devices.

The Fathom can be plugged into any USB-capable device to enable on-board computer vision without needing to make calls to the Cloud. For applications like drones and self-driving cars that require extremely low latency, that’s a good trick.

On the software side, the Myriad Development Kit (MDK) includes a groundbreaking software development framework that enables developers to incorporate proprietary functions of their own and build arbitrary processing pipelines while taking advantage of the optimized software libraries provided. Also included are a rich set of vision, imaging, and linear algebra libraries, plus reference vision processing pipeline examples, all provided as source code. The MDK includes all the programming and debugging tools necessary to enable developers to innovate and differentiate.

Sign up for our insideHPC Newsletter