“Thanks to the large language model to promote the development of artificial intelligence, the research and application of graphics and imagery ushered in new opportunities, is promoting the V7668A-1310B0 development of large language model to large visual model.” On December 30, 2023, on the occasion of the old year and the New Year, Wang Yaonan, academician of the Chinese Academy of Engineering and director of the National Engineering Research Center for Robot Vision Perception and Control Technology, shared the latest trend of intelligent development of machine vision at the 19th Young Scientists Conference of the Chinese Society of Image and Graphics.

In an interview with Nandu reporters, Wang Yaonan said that artificial intelligence makes machine vision intelligent, and higher computing power can support the training of large visual models. V7668A-1310B0 However, smarter large vision models also require higher levels of computing power, better model architecture, and more efficient learning algorithms.

The conference was sponsored by the Chinese Society of Image and Graphics, Pazhou Laboratory, South China University of Technology, Sun Yat-sen University, the Youth Working Committee of the Chinese Society of Image and Graphics.

Talk about industry trends

V7668A-1310B0 From visual computing to visual intelligence

Wang Yaonan: China’s machine vision research has a history of nearly 40 years, initially started from the research of sensors, that is, to convert light information into picture information. The next thing to do is visual processing, including enhancing the obtained image to make it clearer.

Once we have a clear image, we need to get the object we are interested in from the picture. For example, in the field of driverless driving, machine vision needs to detect objects in pictures and answer questions about which are people and which are cars.

These are what we call the three major areas of machine vision. I boil it down to: imaging, processing and understanding.

V7668A-1310B0 Machine vision has moved from visual computing in the past to visual intelligence today. After visual computing is combined with artificial intelligence, the cognitive level has been improved, and the ability to understand the complex environment has been enhanced, and the entire visual intelligence industry has developed rapidly in the last year.

The direction of intelligent development of machine vision is to be widely applied to industrial inspection, intelligent manufacturing and satellite remote sensing.

Talking about machine vision, we must talk about its application and application drive technology development. The Chinese Society of Image and Graphics has 30 special committees, which mainly focus on graphic and image research and serve the national economy. The application scenarios of these technologies are very wide, including industry, agriculture, geographic information systems, remote sensing, land resources, etc.

A person can see the world from birth, and then understand the world, 80% of V7668A-1310B0 the information is acquired by sight. Machine vision is to simulate the human eye, eventually reach the level of the human eye, and in some respects exceed the human eye, to see further and see more clearly.