The 4th China Model Recognition and Computer Vision Conference (PRCV 2021) was held on December 19, 2021 in Zhuhai, Guangdong Province. The fourth Chinese Conference on Pattern Recognition and Computer Vision is co-sponsored by the Chinese Society of Image and Graphics, the Chinese Engineering Intelligence Society, the Chinese Computer Society and the Chinese Society of Automation; in the academic field of artificial intelligence, with the increasing influence of China, PRCV has become the top academic event in the field of pattern recognition and computer vision in China.
01 Jushi Technology shares the road of computer vision innovation practice
Jushi Technology Technology Team deeply participated in this PRCV event, Dr. Luo Changzhi, head of Jushi Technology AI Algorithm Team, went to the meeting to share a special report on the exploration and practice of artificial intelligence in China's high-end manufacturing, and the company's deep learning and computer vision R&D team conducted in-depth exchanges with professors and scholars at the conference site. In addition, an independent enterprise booth (B05) was set up at the conference site to display the latest industrial artificial intelligence technology and product achievements, and conduct in-depth discussions and exchanges with scholars from all walks of life.
On behalf of the technical team of Jushi Technology, Dr. Luo Changzhi shared in depth the landing case of Jushi Technology's deep learning and computer vision technology in China's high-end manufacturing at the PRCV 2021 conference, and introduced the practice path of Jushi Technology in ai landing and product innovation in high-end manufacturing fields such as new energy and semiconductors.
In today's open source era, the latest and most cutting-edge AI algorithms, models, and code in academia are almost at your fingertips. However, they are all scattered and single-point technical solutions designed for closed scenes. Around AI technology, especially combined with the latest academic research results, how to quickly integrate innovative AI algorithm models into industrial common scenarios, the real problems that the industry wants to solve are far more complex and challenging than the problems studied by academia.
Dr. Luo Changzhi, head of the AI algorithm team of Jushi Technology
Around the high-end manufacturing needs such as machine vision and robots, Dr. Luo Changzhi introduced: Jushi Technology's industrial AI product innovation comes from two aspects, on the one hand, it is to keep up with the rhythm of academic innovation, because the AI industry is different from the Internet and other industries, and the progress of academic research is one of the biggest engines for the development of the AI industry; at the same time, for a technology company, providing products that solve customer pain points is the ultimate goal, and engineering thinking is the basis of productization. From data engineering to algorithmic models, to system productization, to system iterative optimization of algorithmic models, every step is crucial.
02 Three Major Matrix series products empower High-end manufacturing in China
At present, Jushi Technology has initially established a strong product matrix. In view of the needs of the semiconductor manufacturing process segment, Jushi Technology has developed a series of Matrix Semi complex visual inspection products, in the two dimensions of innovation and domestic substitution of the technology itself, to achieve a breakthrough in the IC industry, the product has just won the "Product Innovation Award" in the IC field.
For new energy scenarios, the photovoltaic AI quality management and detection system MatrixSolar has served a number of large-scale photovoltaic customers. In terms of software engineering, computer vision, machine learning, and distributed system deployment, Jushi Technology has carried out technological innovation to achieve the delivery acceptance of the industry's first machine reduction industry case, and the product has the ability to realize the unmanned production line workshop.
In the field of robotic AI, based on the machine vision servo and deep reinforcement learning technology developed by Jushi Technology, the large-scale robot AI control system MatrixMath has been applied to a number of heavy machine intelligence scenarios, helping the industry to achieve industrial upgrading from automation, digitalization, and intelligence.