Imaging IoT Image Processing Technologies x Latest IoT/AI Technologies
What is Imaging IoT?
Imaging IoT is defined as a generic term that combines differentiating technologies specifically, device installation technologies for gathering high-quality data from sites (edges) using core technologies cultivated by Konica Minolta with AI platforms that integrate various sensor data and perform advanced recognition and judgment. Our aim is to expand our imaging IoT business by providing optimal solutions to meet requests with respect to various things people “want to see” using imaging IoT technology.
Why is Konica Minolta Working on Imaging IoT?
In the B2B manufacturing sector, it is possible to obtain high-quality data from various sensors with business sites (edges) as starting points, use this data as a base for developing high-quality algorithms, and process real world and cyber data in real time. When doing so, it is important to possess sufficient accuracy to stand up to the realities of the medical and manufacturing industries. Konica Minolta has superior devices for collecting image data from sites (edges), and is in a position to face the challenges encountered by 2 million corporate customers in 150 countries around the world directly. A strength of Konica Minolta’s imaging IoT is that, after accessing both “high-quality data” with sites (edges) as staring points and “working professionals” at said sites, it makes it possible to polish deep learning specialized for image recognition. We will leverage this strength to provide solutions using imaging IoT technology optimized for both hardware and software in “B2B2P” sectors that support professionals at sites (edges).
Top Management Message
Toshiya Eguchi, Executive Officer, Responsible for Technologies, IoT Service Platform Development, Imaging-IoT Solution Business and Visual Solution Business
“Imaging IoT Technology” an Engine for Driving Konica Minolta’s DX Strategy
In recent years, there has been a worldwide trend driving many companies to engage in digital transformation (DX), however, it is said that many of these companies have stepped into this arena without any specific notion of what to do. The “Guidelines for Promotion of Digital Transformations ” announced by the Ministry of Economy, Trade and Industry define DX as “businesses transforming their products, services, and business models based on customer and social needs using data and digital technology in response to drastic changes in business environments, and establishing competitive advantages by transforming how they work, their organizations, processes, corporate traditions, and corporate cultures.” In other words, DX isn’t just creating new businesses using AI or IoT technologies, it’s using these technologies to remake whole companies by transforming everything, including existing products and services, business models, and internal work.
As a “digital company with insight into implicit challenges ,” Konica Minolta aims to promote DX alongside its customers and is working to remain a company the world needs. Konica Minolta’s DX has advanced while strengthening the company’s unique “imaging IoT technology.” We will rise to the challenge of transforming away from a traditional sales business to a service provider with strengths in data use via imaging technology by combining the latest IoT and AI technologies with the “imaging technologies” (optical devices and image sensing technologies) we have developed across all of our business as a strength throughout our long history.
To remain a DX promoting partner customers can rely on well into the future, we intend to strengthen our “imaging IoT technology” and accelerate the promotion of our proprietary DX.
Edge-side AI as a Differentiating Element and Three Application Areas
Konica Minolta is strengthening its “imaging IoT technology” to gain a competitive advantage over other companies via the following three strategies: (1) Develop proprietary “image sensing devices (printed material inspection scanners, ultrasonic probes, spectrophotometers, etc.) (2) Strengthen imaging AI (deep-learning technologies) targets (targets: human behavior, advanced medical care, product inspection) (3) Strengthen “embedded AI” mounting technologies for high-speed processing of AI algorithms (realization of real time processing using edge devices)
Our ability to realize AI processing of large volumes of data on the edge side, which is typically performed on the cloud, in real time with high accuracy is a source of high added value. Because we have 2 million corporate customers in 150 countries around the world and are in a position to face the challenges they encountered directly, Konica Minolta is daily advancing development with the aim of becoming a game changer in the IoT market using platforms we have just created.
We are already leveraging our imaging IoT technology in developing care support solutions, go insight (store marketing), HSTT (cell testing), IQ-501 (industrial printing), and other products and services, and will accelerate the creation of new imaging IoT businesses using this technology as the driving engine going forward.
Imaging IoT Technology
AI Technology Development
Konica Minolta’s imaging IoT AI technology involves the development of the AI technology needed for the three categories of “human behavior,” “advanced medical care,” and “inspection.” These three categories have the special characteristics of being highly consistent with Konica Minolta’s edge strategy and broadly applicable. Since images and video require the learning of large quantities of data, Konica Minolta specializes in these three categories and has created unique learning environments for optimization. Although image and video AI processing may be implemented using the abundant resources available in the cloud, Konica Minolta will leverage its experience performing high-speed image processing for information and medical devices, and its strengths as a manufacturer to achieve low power consumption and mount high-speed AI functions on the edge side (embedded devices, on-premises servers, etc.). In other words, given that having technology that enables end-to-end AI system optimization is a strength, it forms the base of Konica Minolta’s imaging IoT strategy.
Even within AI, deep learning is getting a lot of attention. Deep learning is being used to realize dramatic increases in accuracy in various fields, such as image and voice processing. On the other hand, deep learning calculations require massive computer resources, which is thus increasing the importance of power-saving, high-performance dedicated hardware. Konica Minolta has developed the “NNgen” high-level synthesis compiler jointly with Shinya Takamaeda, an Associate Professor at the University of Tokyo, for easily mounting trained deep learning models on field programmable gate arrays (FPGAs), which it has made open to the public by publishing it as open source software. NNgen enables even engineers and designers who do not have in-depth knowledge of hardware tuning to efficiently develop accelerators that run at high speeds on FPGAs, based on trained deep learning models. As a result, they are able to realize products and services that implement AI processing in real time on the edge side, such as devices and equipment mounted with FPGAs.
Konica Minolta has an edge IoT strategy for developing business in a variety of fields. Therefore, Konica Minolta has developed IoT Platform to serve as a common platform for edge IoT. IoT Platform not only performs on-site analytical processing and cooperate with the cloud as required for edge IoT strategies, but it also provide common functions for meeting non-functional requirements, such as the equipment management and security required when actually installing equipment on-site, which enables the efficient and agile provision of solutions. IoT Platform is made up of three levels (cloud, edge, and device) and is provided in advance with the functions needed for each.
[Cloud] IoT Platform cloud services provide APIs for various operations, such as managing data retention and searching, sending email and mobile push notifications, and managing devices. [Edge] The edge is computers installed at sites and is responsible for functions, such as receiving information from devices, executing processing via deep learning, and sending results to the cloud. [Device] Devices refer to sensors and accelerators installed at sites, and the embedded systems that control them.