News has circulated around social media concerning DFRobot® and its very own Industry AI Developer Contest that brought many great minds together to think up incredible projects which could utilize the LattePanda Delta combined with the Intel® Movidius Neural Compute Stick 2, while also using the OpenVINO™ software to accomplish some truly phenomenal projects.
We have done some work to translate this post for you, which was originally circulating through WeChat and Weibo. When reposting, please refer to the original Chinese article.
"DFRobot® Industry AI Developer Contest" Concludes, Empowering Innovation and Helping Chinese Developer Ecological Construction
Developer Contest Popularity Award: Goodie Box (10841 votes)
- All-in-one virtual sports personal training software (Technical Breakthrough Award)
- Data-driven intelligent grip system (Best Team Award)
- OpenVINO+RoboMaster automatic aiming (Outstanding Contribution Award)
- Industrial inspection drone platform based on OpenVINO (Personal Elite Award)
- RF-Breathing non-contact identification and physiological detection using neural network (Intelligent Creative Award)
- Machine vision-based substation pressure plate status recognition system (Commercial Value Award)
As the "new infrastructure" gradually rises, new technological infrastructure represented by artificial intelligence, cloud computing, and blockchain will become the focus of the next stage of development and construction. Among them, a new round of industrial transformation represented by artificial intelligence, as a core driving force, will promote the transformation and upgrading of trillions of digital economy industries.
The State Council’s "New Generation Artificial Intelligence Development Plan" pointed out that by 2025, the scale of China's core artificial intelligence industry will exceed 400 billion yuan, and the scale of related industries will exceed 5 trillion yuan.
How to use artificial intelligence to achieve industrial upgrading, how to realize the docking of innovative resources, and how to accelerate the implementation of artificial intelligence applications have all become the topics of concern to developers across the country.
On the evening of August 6, DFRobot and Intel jointly organized the "Intel® OpenVINO™ Tool Suite Leader Alliance Assembling Action DFRobot® Industry AI Developer Contest" cloud road show and awards live broadcast through the online platform Bilibili where quintessential application scenarios using artificial intelligence were chosen from those made by developers across the country.
In the end, 6 project teams and individuals won 6 major awards, including the Best Team Award, Individual Elite Award, Technology Breakthrough Award, Intelligent Creativity Award, Commercial Value Award, and Outstanding Contribution Award, and were honored as members of the OpenVINO™ Toolkit Leader Alliance. DFRobot created the title of Lion Program member, thus injecting new vitality into the development and ecological construction of the artificial intelligence industry.
More Than 300 Top Developers Compete for 100 Days
On April 22, 2020, DFRobot and Intel, with the assistance of the DF Maker Community and Mushroom Cloud Maker Space, launched the "Intel® OpenVINO™ Toolkit Leader Alliance Consortium Action DFRobot® Industry AI Developer Competition", which aimed at influencing artificial intelligence industry developers, makers, and business partners to show off quintessential application scenarios from a variety of different fields. This was done to provide developers with a broad communication platform and inject more innovative vitality into the development journey for most makers.
On the evening of August 6, the "Intel® OpenVINO™ Tool Suite Leader Alliance Assembled Action DFRobot® Industry AI Developer Competition" cloud roadshow and awards live broadcast was held online on the Bilibili platform.
At the final stage, Gao Yu, General Manager of Intel China Technology Department, Ye Chen, Founder and CEO of Zhiwei Robot, Zhang Yu, Chief Engineer of Intel and Chief Technology Officer of Intel China Internet of Things Division, Liu Qi, Managing Partner of Tonghao Capital and Founder of Maizi Technology Kai, Huang Zhengqin, Marketing Director of Intel IoTg China, Xia Qing, co-founder of Mushroom Cloud Makerspace, along with a few other guests, the ToT team of the ICIE Lab of Xidian University, the Xingda Tianxia team of Shanghai Dianji University, individual Contestant Li Leheng, individual contestant Chen Qiuyang, the ShanghaiTech University School of Information Femtosecond Intelligent Manufacturing Team, and the North China Electric Power University Smart Energy and Information Research Center team stood out from more than 300 contestants, and won the Best Team Award, Individual Elite Award, and Technology Breakthrough Award, Intelligent Creativity Award, Business Value Award, Outstanding Contribution Award, and six other awards, and also won the titles of OpenVINO™ Toolkit Leader Alliance Member and DFRobot Creative Lion Program Member.
DFRobot, the organizer of the event, said that after 100 days of fierce competition, the competition came to a successful conclusion, injecting new vitality into the current ecological construction of artificial intelligence industry developers.
DFRobot has always regarded Empower Creation as its core mission, and helping the creators of present and future human societies to create better things is what DFRobot deems to be its greatest value.
In terms of empowering developers and helping developers to build an ecosystem, DFRobot and Intel's concepts coincided, which led to each contributing efforts to hold this competition. In the future, DFRobot will continue to work hand-in-hand with Intel and other industry's top companies, institutions, developers, and creators to empower innovation and help the construction of an accessable and cooperation-filled creative open source ecosystem.
Project Roadshow Shows AI Application Imagination Space
In the live broadcast of the awards held that evening, the award-winning developers used an online "cloud roadshow" to showcase the software platform OpenVINO™ toolkit, the hardware platform Intel® Neural Compute Stick 2 and the single board computer LattePanda Delta (Latte Panda), the applications of which which had artificial intelligence projects covering popular areas such as smart industry, smart health and robotics.
Among them, the "All-in-one virtual sports personal education software" that won the Technology Breakthrough Award was developed by the ToT team of the ICIE Laboratory of Xidian University. According to this student team with an average age of 24, the entire virtual private education system analyzes the training and competition videos of relevant athletes in advance, and then collects and records private exercise processes in real time. Through the final data analysis, individuals can adjust their sports performance in time. The method is continuously optimized; afterwards, the establishment of various sports databases and personal comparative learning methods can be realized, and the system will correct personal improper actions in real time.
Guest Huang Zhengqin believes that the highlight of the ToT team's project is to create a relatively universal sports personal training platform, and its technology can be easily migrated to other sports, which has a certain market prospect.
The team from the Center for Smart Energy and Information Research of North China Electric Power University won the Commercial Value Award for its machine vision-based substation pressure plate status recognition system. The team found that the current state of the pressure plate in the substation operation and maintenance work is completely manual, which not only consumes time and labor, but also easily causes visual fatigue, and also leads to misoperations and omissions, so they used the LattePanda+NCS2+OpenVINO-based vision solution which uses the SSD model to develop an artificial intelligence recognition system based on the small size, large number, and dense distribution of the plates. The team improved the accuracy of the system to about 90% by combining the results of multi-frame detections and using the asynchronous mode of OpenVINO, where the delay was controlled within 60ms on the combination of LattePanda + NCS2, which met the basic needs of the system.
Guest Zhang Yu commented that the pressure plate status recognition project of the transformer station provides optimized solutions to real problems, the scenarios in which it could be used is clear, the project can be directly used in production, and the foundation is strong, adding that he hopes to see more and more similar projects that can truly act as vision-based solutions to solve the current difficulties that the industry faces.
Roundtable: Experts Discuss the Future Direction of Developers
In the entries in this competition, in addition to popular keywords such as smart industry, smart health, and smart home, topics such as "smart medical care," whose solutions can solve issues related to the current epidemic, "edge computing," which is closely associated with artificial intelligence applications, and the "construction of the developer community ecology" has also become the focus of heated discussion among the players and guests.
Regarding the concept of edge computing and edge computing intelligence, Dr. Yu Zhang, Chief Technology Officer and Chief Engineer of Intel’s Internet of Things Business Unit in China, explained: “Most of the past artificial intelligence technologies have two extremes, either relying on the powerful computing power of the data center for calculation and reasoning, or using the limited computing power of local equipment to run limited artificial intelligence applications.
Both of these have limitations; that is, the data center approach cannot meet the needs of real-time application scenarios, and the local equipment will have insufficient computing power. In addition, the data storage and aggregation capabilities are insufficient, resulting in problems that can be solved to a very limited extent. In this context, intelligent edge devices with powerful computing power and the ability to respond quickly with low latency on site have emerged. In addition, in Intel's view, intelligent edge devices will be a key technological turning point for realizing the value of data, which is also one of the areas that Intel focuses on."
Ye Chen, the founder and CEO of Zhiwei Robot, believes that edge computing is a computing method that is closer to the data source and has advantages in real-time applications and security. As more and more data and applications increase, edge computing power becomes more and more important. In the software ecology, only when the algorithm model can be calculated locally can the application diversity requirements be realized. "The resulting technology from the combination of LattePanda and VPU has strong reasoning functions. It can collect real-time data from local sensors while meeting the requirements of energy consumption. The small size also allows it to be used in even more application scenarios."
In addition to viewing how artificial intelligence technology plays a role in medical health in the current post-epidemic period, Liu Qikai, managing partner of Tonghao Capital and founder of Maizi Technology, proposed that the epidemic has highlighted the importance of medical services and showed how to use artificial intelligence to improve medical services capacity and quality while maintaining medical services under especially difficult circumstances, and that these are key indicators of technological innovation and medical innovation.
"The medical and health fields are related to life, and there are more things that must be taken into consideration for the application of such technology. I saw many technological ideas in this competition, and we need to find points of innovation stemming from these ideas that can enter into the actual medical environment. In addition to considering technical feasibility, these ideas must also be converted into actual products, which means that they must not conform simply to the current medical service path, but also suit the risk requirements of laws and regulations." He suggested that the combination of artificial intelligence and medical care should find a base threshold at which it could thoroughly integrate into the industrial field, and could then guide the new direction of entrepreneurship and investment.
While looking for technological breakthroughs and industrial incorporation methods, it was discvoered that the cultivation of innovative talents and the ecological construction of developer and maker communities were still the key to the development of the industry.
In this regard, Xia Qing, the co-founder of Mushroom Cloud Maker Space, shared his experience: “Our community construction includes various online and offline activities. We've used a lot of manpower and material resources to build an open online community. Now, the guest community has more than 40,000 active fans and users.
The Mushroom Cloud Maker Space is the core of the offline community, where it can conduct face-to-face communication. Through the construction of online and offline communities, makers and developers are provided with opportunities to learn and communicate, and a large number of tutorials are also provided for developers and makers to learn, just like how the contestants in this competition also have a lot of experience to share, which paves the way for latecomers, speeds up the development process, and unleashes the maker's spirit which can turn dreams into reality. "He said that he hopes to further strengthen cooperation with high-quality communities in the industry such as the Intel Navigator Alliance and jointly contribute to the development of the developer ecosystem."
About OpenVINO™ Toolkit
Intel® Distribution's OpenVINO™ Toolkit Intel® Distribution's OpenVINO™ Toolkit is a free commercial product that provides developers with improved neural network performance on various Intel® processors and helps them further unlock the potentials of cost-effective real-time vision applications. This toolkit supports deep learning inference and easy heterogeneous execution across multiple Intel® platforms (CPU, Intel® integrated graphics, Intel® Movidius™ Neural Compute Stick, Intel® Gaussian Mixture Model-Neural Network Accelerator (Intel® GMM-GNA) and Intel® FPGA) thereby providing implementation across cloud architectures to edge devices. It also provides traditional computer vision libraries (OpenCV *, OpenVX *) and media encoding/decoding functions.
About Intel® Neural Compute Stick 2
The Intel® Neural Compute Stick 2 is supported by the Intel Movidius™ X VPU, which provides industry-leading performance, power and energy efficiency. The Neural Compute Stick supports OpenVINO™, a toolkit that speeds up solution development and simplifies deployment of those solutions. The Neural Compute Stick 2 is easy to use and plug-and-play, and supports common frameworks and sample applications out of the box.
About the LattePanda
LattePanda Delta, the mathematical symbol △, signifies the balance between product design, performance and price, and perfectly reflects the beauty of engineering. The Delta series uses Intel’s new N-series Celeron 4-core processor, which has speeds of up to 2.40 GHz, and has 4GB of RAM natively. It is fully capable of controlling major robots, interactive project cores, IoT edge devices and artificial intelligence brains. Pre-installed with Windows 10 Pro, it also supports a Windows and Linux dual operating system setup.