Page 23 - EETimes Europe June 2021

EE|Times EUROPE 23

Tools Move Up the Value Chain to Take the Mystery Out of Vision AI

in comparison to what PC or cloud developers have and feel like the PROTOTYPE, BENCHMARK, AND TEST AI INFERENCE
unloved stepchild.” IN THE CLOUD
Taking this argument to the next level, he added, “You might have Intel DevCloud for the Edge lets users develop, prototype, benchmark,
embedded systems expertise, but there is a chance that these develop- and test AI inference applications on a range of Intel hardware, includ-
ers may have never worked with image data or deep neural networks.” ing CPUs, integrated GPUs, FPGAs, and vision processing units (VPUs).
Skills are a big challenge, said Bier. “We may have done some spread- With its Jupyter Notebook interface, the platform contains tutori-
sheet calculations and said, ‘Yes, it is possible to run this kind of deep als and examples preloaded with everything required to get up and
neural network at sufficient performance in our application.’ But do we running quickly. This includes pretrained models, sample data, and
know how to do that? Do we have the skills? For most organizations, executable code from the latest version of the Intel distribution of
the answer is no, because they have not had the opportunity to use OpenVINO toolkit, as well as other tools for deep learning. All sup-
this technology in the past. Since it’s a relatively new technology in porting devices are configured for optimal performance and ready for
the commercial world, they don’t have the expertise. They don’t have inference execution.
a machine-learning department or a computer-vision department in The most significant benefit for the developer is that the platform
their company. does not require any hardware setup on the user side. The Jupyter
“In the last couple of years, this has turned into a really big bottle- Notebook’s browser-based development environment enables devel-
neck with respect to commercial application of computer vision and opers to run code from within their browser and to visualize results
deep neural networks — just the know-how,” he added. instantly. This lets them prototype computer-vision solutions in Intel’s
The technology is becoming more accessible, however, as companies cloud environment and watch their code run on any combination of its
have sought to address the skills gap over the past couple of years. “The available hardware resources.
knowledge and skills gap has been quite big but is getting smaller,” said There are three main benefits to this cloud-based offering. First, it
Bier, adding that “a couple of companies,” one big and one small, “have addresses the issue of hardware choice paralysis. Developers can run AI
led the charge on this.” The large company is Intel; the small one is applications remotely on a wide range of hardware to determine which
Edge Impulse. is best for their solution based on factors such as inference execution
“Intel has often impressed me as bucking the trend and willing to time, power consumption, and cost.
make big investments in software tools in a number of ways,” said Bier. Second, it offers immediate remote access to the latest Intel edge
“They have, for example, the OpenVINO tool chain for edge computer hardware. On the software side, it addresses the issue of having to
vision and inference, and DevCloud for the Edge. Edge Impulse is also
a cloud-based environment. To an embedded developer, [the cloud
environment] feels weird. Everything to them is often on their desk —
the dev board, the workstation, the tools — and they don’t even need an
internet connection. Everything is very local. So it feels very strange to
say, ‘Put your code in the cloud’ and run the tools in the cloud.”
The trend addresses time to deployment as well as the skills gap.
A frequent frustration for embedded developers is getting access to
boards and tools and getting them properly installed, Bier said. The
timeline is “usually measured in weeks, sometimes in months, and
that’s painful, especially if at the end of that, you realize that’s not
what you needed and you need to repeat the process with some other
boards.” For example, you might find at the end of the process that
“you need the next one up, with higher performance or a different set
of I/O interfaces.”
But if the supplier “has all the development boards in the cloud
connected to their machines and [can] access them at will, that offers
tremendous convenience. Likewise, they have got the latest versions of
the software tools, and they’ve sorted out all the dependencies between
them.”

PAVING THE WAY FOR IMPLEMENTING VISION
So how do you speed up the deployment of embedded vision to enable
features such as object detection and analysis, whether for smart cities,
factories, retail, or any other application?
Having realized the pain points described by Bier, companies are
addressing them. Some now offer tools such as cloud-based devel-
opment systems that allow you to feed your code or data and get
evaluations in next to no time. Others provide reference designs that
allow you simply to plug in your camera output and choose from librar-
ies or apps that provide inference algorithms for common applications.
In the former camp, Intel DevCloud for the Edge and Edge Impulse
offer cloud-based platforms that take away most of the pain points
with easy access to the latest tools and software. In the latter, Xilinx
and others have started offering complete systems-on-module with
production-ready applications that can be deployed with tools at a A trained model showing predicted on-device performance
higher level of abstraction, removing the need for some of the more estimations in Edge Impulse’s web-based user interface
specialist skills. (Source: Edge Impulse)

www.eetimes.eu | JUNE 2021