Software Simplicity: The Solution to Vision at the Edge

February 6, 2019 Jeffrey S. McVeigh, Intel

This is the first of three blogs from Intel that describe a developer’s journey toward designing an AI-based system. The second blog covers The Journey from Prototype to Production for AI at the Edge.

Deploying vision capabilities on edge platforms requires difficult tradeoffs between latency, throughput, memory footprint, communication bandwidth, power and cost. Luckily, there’s an ever-growing diversity of hardware choices to allow system designers to select the best option that meets their needs. Historically, hardware diversity also implied time-consuming software development to port vision applications and algorithms to a new hardware target and optimize for real-time constraints. Intel is working with the industry to solve the puzzle of hardware diversity for traditional and deep learning based vision at the edge.

The Intel® Distribution of OpenVINO™ toolkit (which stands for Open Visual Inference and Neural Network Optimization) enables developers to streamline the deployment of deep learning inference and high-performance computer vision applications across a wide range of vertical use cases at the edge. The toolkit is compatible with popular open source deep learning frameworks, and enables developers to easily target execution on CPUs and accelerators (GPU, FPGA, VPU) specially designed for AI inference, such as the Intel Vision Accelerator Design Products. The beauty of the toolkit is that it provides a unified and common abstraction for AI inference across diverse hardware targets with a comprehensive and intuitive API, merging simplicity with optimal performance. Software simplicity and performance – just what the developer ordered!

The toolkit has been used by numerous industry-leading companies and innovators to deploy AI and vision commercial solutions across numerous segments, including healthcare, surveillance, smart cities, retail, robotics, and web services. Check out a few case studies here. For example, developers have used the toolkit to build applications that automatically track movement of vehicles and pedestrians, detect manufacturing anomalies, aid doctors in medical diagnoses, monitor the awareness state of machine operators, and analyze the reaction of customers at retail product placements.

The Intel® System Studio, an all-in-one cross-platform tool suite, helps developers accelerate system bring-up, boost system and application performance, power efficiency, and reliability. It consists of a high-performance compiler and a rich set of libraries, debuggers, and analyzers that complement the AI inference and vision capabilities of the Intel Distribution of OpenVINO toolkit.

Both the Intel Distribution of OpenVINO toolkit and Intel System Studio are pre-installed on the IEI Tank* AIoT Developer Kit to provide developers a quick path for AI development. Along with the software tools, the hardware kit contains a powerful Intel Core processor with integrated HD graphics, ruggedized chassis, rich set of I/O ports, drivers, and sample code. The developer kit can be used for both prototyping and deployments.

These solutions can be mixed and matched with other Intel technologies, tools, or external systems to develop situation-specific software to improve the performance of time-critical tasks in wireless communication, develop security and surveillance solutions, and scale high-quality and low-cost video conferencing.

As Leonardo da Vinci once said—“Simplicity is the ultimate sophistication.” Visit Intel at the 2019 Embedded World to experience the renaissance of software simplicity and how it can help speed your development of vision at the edge.

Intersecting AI, Vision & Workload Consolidation at the Edge

Intel @embedded_world

Experience advanced technologies putting data to work from edge to cloud. Deliver high performance AI with broad connectivity and security. Connect with Intel at Embedded World on Feb. 26-28 in Nuremberg, Germany. See live, eye-catching demos that detect cancer cells in unseen images, manufacturing defects in objects on a conveyer belt, and many more—built using integrated hardware/software solutions that make a developer’s job much easier.

Visit Intel booth Hall 1-338 | Attend an Intel-led tech session

More Resources

*OpenVINO and the OpenVINO logo are trademarks of Intel Corporation in the U.S. and/or other countries.

Previous Article
ADACORE, NVIDIA Accelerate ISO 26262 Compliance by Securing Critical Firmware with Ada and SPARK

NVIDIA will also migrate certain SoC product lines to the RISC-V instruction set. The hope is both moves wi...

Next Article
Save energy and retrieve performance with RISC
Save energy and retrieve performance with RISC

Compact PICO-ITX board with RSIC CPU for KIOSK and POS applications combining low power consumption and hig...