New Device Brings Computer Vision to IoT and Edge Computing

Neural Compute Stick 2 is a new device created by Intel. / Photo by: Wojciech Balsewicz via 123rf


Intel has been building supporting software and technology for new age smart devices. Recently, the company has revealed that their new product will make it much easier to create smarter artificial intelligence (AI) algorithms and computer vision applications at the network edge. This device, called the Neural Compute Stick 2 (NCS 2), may look like a regular USB drive, but it can do so much more than just store data on different devices.

Whenever a computer vision application based on neural networks is run, the computations are then offloaded onto a special chip that is located inside the NCS 2 stick, lessening its load and compartmentalizing the work needed to make it function. Its primary function is to accelerate the computation necessary to make executing neural networks much faster when compared to using a standard GPU. This device acts as a micro GPU in that regard for inferencing machine learning models. Once trained, this machine learning model could be used in a plethora of smart and Internet of Things (IoT) enabled devices to more efficiently manage the processing inside and to control it more smoothly. This can then learn and understand different objects around it, and classify them into different parts and segments for future use. The NCS 2 is most helpful when creating devices, and improving upon and studying a service as a whole. The main difference between this device and other similar gadgets run like this is that the NCS 2 is affordable and cheaper than most.

This device can allow developers to go from prototyping to production leveraging a range of Intel vision accelerator form factors into use in the real world. While the first generation of the NCS series was launched last year, the NCS 2 comes with a whole new set of abilities at a reasonable price.