Sharing collaboration examples such as Intel Movidius VPU
Discussion on CCTV, 3D cameras, data management plans, etc.

A meeting was held to discuss AI edge solutions essential for 3D cameras, video security, and smart city data management, targeting the global IoT market.
Intel Korea held the Korea Edge AI Forum at the InterContinental Seoul Coex on December 3 to share Intel's next-generation AI vision and innovation strategy.

The event was attended by Intel Korea President Kwon Myung-sook, LG Electronics Executive Vice President Ji Seok-man, Hanwha Techwin Executive Vice President Jung Won-seok, Innodeb CEO Lee Seong-jin, and Intel Korea Executive Vice President Choi Byeong-won, who presented case studies that help solve the AI challenges currently facing companies.


Edge AI solution that maximizes data analysis and utilization
Intel Korea President Kwon Myung-sook presented a total of eight industry-specific opportunities that Intel is currently focusing on under the theme of 'Intel Edge AI Strategy': smart cities, financial services, industrial, gaming, transportation, home and retail, robotics, and drones.>
In addition, Intel introduced its edge AI solutions designed to realize safe and efficient household, industrial, and public environments by recognizing human faces and distinguishing a wider range of objects.
He explained that to maximize data analysis and utilization, computing innovation and the adoption of AI technology at the edge, where the most data is generated, are necessary, and stated that Intel is providing platform solutions encompassing both hardware and software to achieve this.

Intel added that, particularly for edge AI solutions, creating new use cases by leveraging technology is crucial, and therefore, the company is actively working with domestic partners to identify business models for ecosystem building and technology commercialization.

Intel is supporting products specialized for endpoints, edge, and data centers to increase the utilization of AI at the growing edge. In 2019, 45% of data was analyzed and utilized at the edge, and it is projected that 43% of AI workloads will be processed at the edge by 2023.

Intel supports users in easily optimizing AI algorithms and efficiently commercializing them by providing the Intel Open VINO Toolkit, which is applicable to a diverse range of processor families, from VPUs to Xeon platforms for data centers.

He stated in the announcement, “We plan to launch the Intel Movidius VPU in 2020,” adding, “It will increase product efficiency and help users accelerate AI implementation.”

To this end, Intel is building a solution partner ecosystem that includes hardware, software, system integration, solution integration, and sales channels, and supports the entire process through a collaboration model to enable customers to develop and launch AI-based products.


Video intelligence 3D camera for facial recognition and body measurement
In the second keynote speech, Senior Vice President Ji Seok-man of LG Electronics presented on the current status of AI technology development being pursued in collaboration with Intel, under the theme of 'Camera and Image Intelligence'.

LG Electronics is developing a 3D camera capable of recognizing the distance between the camera and the subject in HD resolution, and is implementing application products such as ThinQ Fit (virtual clothing fitting solution) by combining it with deep learning technology in areas such as face recognition, body measurement, and obstacle recognition.
Cameras are recently transitioning from an era of image quality to one of recognition, and LG Electronics is developing a 3D camera to respond to this. To create products capable of generating various forms of business in the B2C market, the company has adopted Intel RealSense camera technology.

LG Electronics aims to create high-performance 3D cameras by combining Intel RealSense's 3D depth sensing technology with its existing optics and applications. It is currently conducting research and development to enable 3D cameras to perform roles in home appliances, robots, and other applications that were previously impossible with 2D technology.

To provide these features through an open strategy, LG Electronics also introduced Intel's D4 chip.


Realizing Incident Prediction and Real-Time Response with Deep Learning CCTV
Hanwha Techwin Executive Director Won-seok Jeong stated under the theme of "Video Security and AI Technology Based on Intel Solutions" that the company is adopting Intel's processors as well as the Intel OpenVINO toolkit, which optimizes hardware, to perform high-definition compression, transmission, and intelligent analysis—the core of video security.

Since CCTV must now enable real-time response rather than post-mortem analysis, and must evolve to allow for proactive prediction and prevention in the future, the introduction of AI technology at the edge is essential. We are applying OpenVINO toolkit and VAS (Video Analytics Suite)-based solutions to optimize software algorithms for use at the edge.
Currently, Hanwha Techwin has adopted Intel's microprocessors and installed them in server-type NVRs (Network Video Recorders), and plans to launch NVPR in 2020 with deep learning-based video analysis artificial intelligence technology.


Data Management and Extraction Platform, Key Element for Smart City Implementation
Lee Seong-jin, CEO of Innodeb, stated under the theme of "Smart City Data Management Platform Based on Intel Edge AI Solutions" that the company is collaborating with Intel to develop a city control and data management platform necessary for the implementation of smart cities.

He argued that image quality is no longer the standard for CCTV performance and explained that the technological capability to extract meaningful data from collected footage is becoming the key. To this end, technological evolution is required across the entire spectrum, and key solutions, including Innodeb's Data Management Solution (DMS), are equipped with Intel's latest Xeon processors.
Recently, edge solutions applying AI technology based on the Intel Vision Accelerator Design and Intel Open VINO Toolkit, based on the Intel Movidius VPU, are also being developed.


Intel Movidius VPU, Edge AI, and Independent SoC Available
At the tech session, Intel Korea Executive Director Choi Byeong-won introduced the new 3rd generation Intel Movidius VPU, scheduled for release in 2020, under the theme of 'Next-generation Intel Edge AI Platform'.

Unlike the second generation, this VPU can be used as an independent SoC and can also be installed in AI accelerators, standalone smart cameras, drones, robots, etc. Not only is it more than 10 times more efficient than existing products without consuming power, but it also supports the Intel OpenVINO Toolkit, enabling the application of edge AI.

Intel's hardware can be tested in cloud environments with the OpenVINO toolkit, an AI development and optimization tool, and DevCloud supports the implementation of AI technology with Intel's hardware (CPUs of various performance levels such as Xeon, Core, and Atom, as well as FPGAs and vision processors).
Meanwhile, during the panel discussion, Intel Korea Executive Vice President Park Sung-min served as moderator and shared corporate AI strategies and future AI visions through collaboration with Intel with Intel Korea President Kwon Myung-sook, LG Electronics Senior Vice President Ji Seok-man, Hanwha Techwin Senior Vice President Jung Won-seok, and Innodeb CEO Lee Sung-jin.