MOJO.Think #CEOview : AI at the Edge of networks in IoT devices will be the biggest Taiwan AItech industry. Like Intel, all major IT players are trying to establish themselves as the “standard”. Efficient chip design, smart cloud compute management and customer-centred apps will be key. MOJO is there.

On January 12, 2022, Global Digital MOJO attended the ” Intel: Taiwan Edge to Cloud Conference 英特爾邊緣到雲端技術產業論壇” held at Taipei Nangang Exhibition Center Hall 2.

This event brought together more than 1,000 attendees to discuss Intel’s efforts to push the four superpowers, which are ubiquitous compute, cloud-to-edge infrastructure, pervasive connectivity, and artificial intelligence. There are four main themes represented at this event: 5G Network, Artificial Intelligence, Edge AI/Computing and Cloud Service/Computing. There are 35 topics and sections, ranging from hardware and software to smart cities and AIoT applications.

In the exhibition area were more than 40+ industrial partners showcasing their latest development in edge devices, Smart Factory (ADLINK 凌華/NexCOBOT 創博/Axiomtek 艾訊/Moxa 四零四), Robotic Machinery (Techman Robot 達明/NEXCOM 新漢智能) Intelligent Retail (Advantech 研華/IEI 威強電/Shuttle 浩鑫/LIPS 立普思/PROX 南京資訊), Biotech & Medical Tech (NCKU 成功大學/Qisda 佳世達/Avalue 安勤/AOpen 建碁), AI Recognition/Detection Tech and Edge-AI Solution (LILIN 利凌企業/7Starlake 喜門史塔雷克/Supermicro 美超微/Gigabyte 技嘉/D8AI 八維智能/ASUS 華碩/Wiwynn 緯穎/Everfocus 慧友電/CyberLink 訊連/YUAN 聰泰/AAEON 研揚/IBASE 廣積/Alpha 明泰), Energy solution (YES Energy 裕電能源), 5G & Cloud Solution (QCT 雲達/PEGATRON 和碩聯合/MiTAC 神雲/Lanner 立端/工研院/AEWIN 其陽/Accton 智邦/Microsoft 微軟/QNAP 威聯通/Atrust 冠信電腦/Portwell 瑞傳) and many more.

A keynote speech was presented by CEO Pat Gelsinger, Sandra Rivera Executive Vice President and General Manager of the Datacenter and AI Group, and Nick McKeown Senior Vice President and General Manager of the Network and Edge Group (NEX). To share Intel’s vision and mission in a fast-changing technology world and the transformations underway – from cloud to edge. They presented a research analysis. What edge computing means for infrastructure and operations leaders, Gartner, October 3, 2018: “Massive growth of data with increasingly diverse workload requirements. 75% of data will be created outside the data centre by 2025.”

Chairman of the National Development Council Kung Ming-hsin (龔明鑫) took the stage to present Taiwan’s post-pandemic digital transformation strategy. They identified 5 main catalysts of the 3Ds’ transformation: De-carbonisation, Digital Transformation, Distributed Supply Chain. Impacts include climate change, inflation, pandemics, supply chain destruction, and China-US trade wars. Developing the next Asia Silicon Valley through extensive investment in AIoT applications, startup unicorn ecosystems, merging technologies/systems and increasing outbound capability.

In the showcase, we witnessed the abilities of AI functions used in Smart City with Computer Vision. Such as Traffic Flow Detection, Object/Licence Plate Recognition, Vehicle Type Detection, Vehicle/Pedestrian Auto-count, Movement Tracking.

A complete solution for smart city monitoring hardware. AAEON presents a Smart Pole in the Smart City Blueprint section. Smart Pole is an all-in-one device that can detect the environment, act in an emergency, detect weather, control lighting, show public information, provide wireless internet connection, and provide first-aid kits. The data is processed in real-time on an edge machine before being sent back to the situation room. With OpenVINO, Intel collaborated with AAEON, Arm’s Pelion platform was used to manage devices and data is stored on Microsoft’s Azure cloud-based infrastructure.

AMR (Autonomous Mobile Robot) devices from ADLINK and NexAIot help companies make warehouses more productive and manufacturing more efficient via Industrial IoT. With LiDAR, Ultrasonic sensors, camera to build environment awareness, and auto-docking wireless charging, the NexMOV module gave us a taste of what the future could be like. ADLINK’s compute technologies enabled them to scan objects with artificial intelligence to improve production workflow and increase work safety, as well as using No/Low-Code tools to make building in-house SI more convenient and visualising pipeline control.

Advantech offers a quick solution for the improvement of storage environment monitoring systems in medical facilities, retail stores, food processing centres or normal offices. In the research conducted on the ground and with clients, it was found that the methods were still manual and very inconsistent. The LoRa sensor claims to be able to attach to a variety of surfaces and connecting to Edge-AI gateway for model management and data processing, one can set up functions such as defroster detection, temperature alarms, and automatic report generation. 

To be differentiated from industrial robots, Techman Robot introduces a new cobot (collaborative robot). Cobots are designed to be light-weighted and more flexible in setups. TMvision is a feature which detects barcodes, learns objects, and amplifies image recognition. Vision allows cobots to break through traditional workflow design and identify object locations. The TMLandmark system updates geographical information instantly when objects and the robot’s position or distances change.

“Face recognition is just the tip of the AI iceberg” Private Banker International, shared by Cyberlink Marketing Director Steven Lien in his Edge computing presentation. FaceView, a new application of Advantech that integrates with FaceMe’s self-developed SDK, identifies faces, even with masks on, and is now seeking adoption in retail stores, public transportation, and security during this pandemic. Taking the health add-on a step further, Steven described how it helps detect body temperature and HR features for office check-ins and outs, and how it works with the door lock and access management. AI-detection can be used to safeguard the environment and enable a safer environment.

We see ASUSCloud’s collaboration with the National Taiwan University Hospital (NTUH) to develop a procedure simplification solution to resolve crowded emergency departments (EDs) and improve patient safety. AI diagnoses through audiovisual, historical record, X-ray examination, and physical monitoring of patients. This helps the responding medical team to make timely decisions on the treatment approach for each individual patient. Deep hardware integration, smart data visualisation, and the seamless information transmission from medical devices to personal health tracking devices all through cloud services.

Qisda is building a federated learning framework for Medical-AI applications in collaboration with National Cheng Kung University (NCKU). Training with Intel’s OpenVINO toolkit, using deep learning and computer vision to help the doctors identify lung nodules in 400+ X-Ray images. Categorising a patient’s results as Benign, Probably Benign, Probably Suspicious, or Suspicious. Traditionally, this is a laborious, manual process. This now enables the medical team to focus on other emergencies and have a better chance of segmenting the images after they are processed. A motion-detecting application has also been developed for the Pulmonary Rehabilitation using Edge Client Computing.

Five aspects are to be considered in the use of 5G technology and Computer Vision in Smart Cities: Law enforcement, Public safety, Smart City IoT, and Business intelligence. 7Starlake (喜門史塔雷克) is dedicated to the development of AI-driven hardware integration for smart cities, assisting in the development of smart mobility and increasing safety. In the 5G era, AI-computer vision is transitioning from passively analysing events and reporting errors with Edge AI to analysing live data and calculating motions with a combination of Cloud BI tech. In public transportation such as the bus service, V2X can allow Edge AI to assist the driver on multiple areas, such as traffic flow analysis, alarming dangerous violation activities, detecting the interior space and passenger safety of the car, monitoring the blind spots of large vehicles, and tracking fatigue & distraction behaviour in the driver.

Likewise, DFI is focusing its technology on helping law enforcement. There are plenty of CCTV cameras in every corner of the street and on the road now to record and monitor the safety of citizens. However, when it comes to law enforcement, the speed cameras used by the traffic police are still extremely expensive to install and maintain. In order to take advantage of current hardware (CCTV availability) and use of powerful Edge-AI processing capabilities, DFI worked with the Police Department to develop POCs on AI car plate segmentation and recognition. Streaming live data to a police database to enable law enforcement to take action immediately upon any violation events. This is estimated to save up to 88.6% in maintenance fees and to decrease traffic accidents by at least 30%, to say nothing of the fact that the range of law enforcement will be expanded considerably.

MOJO Think Takeaways

To sum up this event. There is a growing number of Edge-AI solutions that can help to rapidly process real-time data. They could improve city safety, factory production, and various monitoring processes. Redefining the ways we interpret/share/gather information and data, as well as machine-to-machine bridging and communication.

The communication between human and machine is closer to being achieved with computer vision. Traffic flow and transportation types can be identified with Intelligent Traffic Management (ITM) with AI. Spotting suspicious movements and initiating a process with law enforcement regarding illegal activities encountered. AI-powered Medical assistant application to reduce workload for labour intensive image diagnosis.

With the emergence of 5G, hardware capacity, cloud services, and power storage will become more stable and mature. Enabling the era of live-virtuality immersive integration in the metaverse while accelerating Industrial 5.0 with AI mass customisation and cyber-physical cognitive systems. Yet, those technologies are also becoming more human-centric and human-machine collaborative work settings.

Global Digital MOJO has been assisting businesses to integrate advanced processes with artificial intelligence and improve their digital experience with UX design. We are happy to be invited to the Intel conference and be a part of this transformation.