Workpad Industrial Data Solutions IoT
VisionX Leverages Blockchain and AI for Inspection Intelligence
Posted:
Source: Vision X
In early October, Dr. Haisong Gu, co-founder of VisionX (www.visionx.org), presented his latest project to the state-owned enterprise China Coal Technology Engineering Group (CCTEG) in Changzhou and established a strategic collaboration. Within the next three years, the current system of manual inspection at CCTEG could be entirely replaced by cameras and sensors with AI visual inspection technology developed by VisionX.

Dr. Haisong Gu was among the first group of artificial intelligence post-­graduate students incubated by China who were admitted to the Southeast University Computer Vision Research Lab in 1985. In 1993, Dr. Gu obtained his PhD in Computer Vision from Osaka University, a globally renowned institute in the intelligent recognition field. He has also worked as professor and researcher at Southeast University, Osaka University, University of Southern California, PRI and Stanford University; as well as having worked as Senior Researcher at Panasonic and many other esteemed companies in the field of image recognition.

During his 30+ years of involvement in AI application development, Dr. Gu developed the first iteration of an image-­assisted smart training system for China's national diving team in the 90s, using the 'Positioning method' theory, and was awarded first place in the Culture Department's Technology Achievements Award. Utilizing his innovative Computer Vision technology, Dr. Gu generated an automatic scoring system for various applications in gymnastics competition. Dr. Gu also assisted in the construction of the only ‘train ferry' in Chang Jiang which allowed ferry navigation in poor weather conditions and enabled year-round functionality.

In the early 90s, Dr. Gu used machine learning and motion modeling technology to devise an innovative image-­semantic segmentation theory based on MDL. Dr. Gu was invited to publish a paper at the Conference on Computer Vision and Pattern Recognition, regarded worldwide as the most prestigious and acclaimed artificial intelligence conference, for two consecutive years. In Matsushita, the company successfully developed an intelligent robot with three­dimensional object sensing capability, and realized one of the first applications of artificial intelligence, defect sorting and placement, in production lines, and was subsequently awarded the Japan Automation Association Technology Progress Award.

After traveling to the United States in 2000, Dr. Gu developed an operation control system based on gesture recognition at the USC Artificial Intelligence Institute in Nankai. As part of the face recognition startup Eyematic (acquired by Google in 2007, becoming the core of Google's face recognition technology), it successfully commercialized an animation system based on face recognition and tracking. Dr. Gu lead development and was first authour of a research project by Ford which was developed to realize facial expression recognition (CVPR2002) through real-­time tracking of facial feature points and Efman facial motion coding (FACS), and subsequently developed a driver fatigue intelligent detection and control system based on facial expression recognition. As part of the intelligent video surveillance technology startup, Vidient, the technology was used to develop a number of practical algorithms for object and motion recognition analysis based on intelligent video analysis, which was selected by the US aviation department and successfully implemented at the San Francisco International Airport, one of the first applications of such technology in the world. At present, SmartCatch intelligent video monitoring software products have been installed in airports and other transportation hubs around the world.

In 2008, Dr. Gu was invited to design the "Intelligent Real­-Time Passenger Flow Statistics System" for the Shanghai World Expo, as well as the passenger flow management system for the China Mobile Business Office. The passenger flow intelligent statistics implemented at the CCP's first­-level conference site successfully solved issues related to audience management.

Dr. Gu is currently the leader of the AI team at the Silicon Valley World's Top 500 R&D Center. He led the AI R&D team to successfully develop the HSTT image diagnosis system based on deep learning. Utilizing AlphaGo technology in the field of medical diagnosis, recognition accuracy has drastically improved and outclasses current professionals. This innovation was named one of Japan's 2020 core technologies, predicted to have huge applications into the future. It was successfully employed to assist Midea in the realization of AI applications for surface quality inspection of home appliances. It proposed the solution of an intelligent collaborative robot (AI Cobot) to improve flexibility of welding and assembly production of home appliances, and promoted the integration of intelligent manufacturing technology in the industry.

Dr. Gu has published more than 30 academic papers for world-class AI conferences and publications including the Conference on Computer Vision and Pattern Recognition and also the Pattern Analysis and Machine Intelligence journal. Dr. Gu currently hold 20 granted invention patents in the United States, Japan and China and has a further 17 applications pending. He is currently a senior member of IEEE and was a corporate representative of the IEEE International Industrial Standards Technology Association (IEEE­ISTO). On behalf of Konica Minolta and Stanford University, Dr. Gu led a number of industrial joint research projects.

According to available statistics, there are about 60 million quality inspectors engaged in surface defect testing in various industries in China alone. They use the naked eye to do monotonous and repetitive work every day. This is a serious financial and logistical burden among multiple industries. By utilizing its innovative visual inspection technology, VisionX can alleviate this work-load and improve products with increased accuracy and efficiency. By improving quality and speed of production, this technology will eventually replace the necessity for manual labor and potentially reduce production costs by billions of dollars per year. Dr. Gu said, "Using the blockchain's distributed AI technology and model sharing mechanism, we can make the production of individual industries highly efficient. Utilization of our technology has the potential make a great contribution to China's ultimate realization of the national smart manufacturing 2025 strategy."