Shulou(Shulou.com)06/02 Report--

Introduction: China computer Congress CNCC 2019 (10.1719) is about to open in Suzhou.

Editor's note: China computer Congress CNCC 2019 is about to open in Suzhou. It is estimated that more than 8000 people will attend this meeting, including conference reports of 16 well-known experts and entrepreneurs in the computer field at home and abroad, three conference theme forums, more than 70 cutting-edge technology forums, 20 special events, and 100 exhibitions of scientific and technological achievements.

On the afternoon of the 19th, in order to explore the development of computer vision and smart cities, Shen Shengmei and other famous scholars will jointly discuss the cutting-edge technologies of computer vision, such as scene interpretation, face recognition, human posture, and visual content understanding from an academic and industrial point of view. as well as the research achievements, development prospects and technical applications in intelligent security, self-driving and other intelligent city construction are worth looking forward to.

As the hottest research and application direction in the field of AI, computer vision technology is at the forefront of the development of artificial intelligence in academic research and industrial landing. From the popularization of the concept a few years ago to the landing in the field of smart cities, computer vision technology is profoundly changing the concept and way of using information resources in the whole society.

On the one hand, with the research progress of the theories and methods of computer vision and pattern recognition, especially the great success of the new generation of artificial intelligence theory represented by deep learning in the research of computer vision and pattern recognition, scholars continue to put forward a large number of algorithms with their own characteristics and facing various applications, which have greatly promoted the development of related technologies. On the other hand, the development potential of the demand trend computer vision industry is huge, the application scene expands and permeates various industries, and people pursue the improvement of life safety and production efficiency, which determines that the computer vision industry has room for development. At present, visual intelligence technology has landed in financial, security, retail, education, medical, autopilot and other urban scenes and people's lives.

Smart cities have been written into the national strategy, and the construction of new smart cities has entered the fast lane. Computer vision technology can empower the applications and service systems of public safety, urban governance, transportation, community, education, health care, industrial Internet and other industries in smart cities, which is the perception system of the whole smart city. Therefore, how to combine the scientific research of computer vision technology with the construction of smart city is a topic that needs to be discussed by both academia and industry.

In order to explore the research achievements and development prospects of computer vision technology and accelerate the landing of cutting-edge technologies in the field of smart cities, sponsored by the Chinese computer Society (CCF) and undertaken by the management committee of Suzhou Industrial Park, the 2019 China computer Conference (CNCC 2019) held in Suzhou Jinjihu International Conference Center on October 17-19 specially organized the forum "computer Vision Technology enabling Smart City". It will carry out in-depth discussion around the hottest topics in the current technology and industry.

China computer Congress, founded in 2003, is the largest and highest-standard academic, technological and industrial event in the field of computing in China, and has been successfully held for 15 years.

Previous meetings brought together Turing Award winners, experts and representatives of ACM, IEEE-CS, IPSJ, KIISE and other international organizations, as well as domestic authoritative experts, scholars and well-known enterprise technology experts in the field of computing. The scale of this year's meeting is expected to exceed 8000.

This year, with the theme of "Intelligence + leading Social Development (AI+ Leading the Development ofSociety)", combined with the achievements and challenges in various fields at this stage, more than 70 cutting-edge technology forums will be arranged, covering cutting-edge academic, technological, educational and industrial topics such as artificial intelligence, big data, software engineering, computer education, brain-like computing, and health care big data, in order to promote the better and faster development of China's computer industry.

CNCC2019 computer Vision Technology enabling Smart City Technology Forum will focus on the cutting-edge technologies of computer vision technology, such as scene interpretation, face recognition, human posture, visual content understanding, as well as the research results, development prospects and technical applications in intelligent security, self-driving and other intelligent city construction, and conduct in-depth discussions from the academic and industrial point of view.

Shen Shengmei, chief scientist of Pense Science and Technology and president of Singapore Research Institute, served as the main chair of the forum, Chen Xilin, researcher of Institute of Computing Technology of Chinese Academy of Sciences and fellow of IEEE Fellow, IAPR Fellow and CCF, together with Jiao Licheng, professor and doctoral supervisor of Xi'an University of Electronic Science and Technology, Li Ziqing, lecturer and IEEE Fellow of West Lake University, Director Alibaba Autonomous driving lab and chief scientist Wang Gang. Feng Jiashi, Assistant Professor of the National University of Singapore and head of the Machine Learning and Vision Laboratory of the National University of Singapore, Lu Jiwen, Associate Professor of Automation Department of Tsinghua University, and Lao Shiyi, Vice President of Shangtang Technology, General Manager of Intelligent driving Business and General Manager of Shangtang Japan, jointly discussed the development and application of computer vision technology.

Guests of the forum and the topics of the speech are as follows:

The host of the forum

▼

Shen Shengmei

Chief scientist of Pennsylvania Science and Technology and director of Singapore Research Institute. Master of Electronic Information Engineering, Department of Electronic Engineering, Xi'an University of Electronic Science and Technology. Shen Shengmei is a leader in the field of artificial intelligence and deep learning, former vice president of the Singapore (Panasonic) Research Institute, leading an algorithm research team of more than 40 people and accumulating more than 300 patents. It has the full stack technical ability of computer vision, and the technical area spans many application fields. the world's top achievements have been achieved in the fields of face detection and recognition based on deep learning, pedestrian detection and tracking, pedestrian re-recognition, vehicle recognition, autopilot, driver behavior detection, mobile operation robot and so on. In March 2019, he announced to join the domestic artificial intelligence company Pensi Technology as chief scientist and director of the Singapore Research Institute, dedicated to monitoring and safety, smart city, self-driving, intelligent robot and AI factory automation and other related technology research.

Common chair

▼

Chen Xilin

Research fellow, Institute of Computing Technology, Chinese Academy of Sciences, member of IEEE Fellow, IAPR Fellow, CCF. In recent years, the main research fields are computer vision, pattern recognition, multimedia technology and multi-mode man-machine interface. At present it is IEEE Trans. Associate Editor of on Multimedia, Senior Associate Editor of Journal of Visual Communication and Image Representation, deputy editor of Journal of computer Science, deputy editor of artificial intelligence and pattern recognition. It has successively won one second prize in national natural science and four second prize in national scientific and technological progress, co-published one monograph, and published more than 300 papers in important journals and conferences at home and abroad.

Brief introduction of speakers and reports

▼

01.

Jiao Licheng

Complex scene interpretation based on Deep Learning

▲ Jiao Li Cheng

Summary of the report

Scene interpretation is an important problem in the field of computer vision. In many practical image / image processing applications, such as self-driving, augmented reality, battlefield situation awareness and so on, higher performance requirements are put forward for scene interpretation and target recognition. The traditional processing technology based on "feature engineering" has been difficult to meet the existing needs. The mechanisms of human brain information processing, such as remarkable attention, hierarchical perception and sparse perception, bring new opportunities to complex scene interpretation and target recognition. This report focuses on deep learning techniques and shares the team's research progress in scene interpretation and target recognition.

A brief introduction to the reporter

Jiao Licheng is a professor and doctoral supervisor of Xi'an University of Electronic Science and Technology. He is currently the Director of the Department of computer Science and Technology of Xi'an University of Electronic Science and Technology, the Director of the key Laboratory of the Ministry of Education for Intelligent perception and Image understanding, the Director of the International Joint Research Center for Intelligent perception and Computing, the Director of the Joint Laboratory for International Cooperation in Intelligent Information processing ("111 Program"), and the member of the Science and Technology Commission of the Ministry of Education. Expert Group of artificial Intelligence Science and Technology Innovation expert Group of Ministry of Education, Vice President of China artificial Intelligence Society, Vice Chairman of National University artificial Intelligence and big data Innovation Alliance, main seat of IET Xi'an Branch, main seat of IEEE Xi'an Branch Award Committee, main seat of IEEE Computational Intelligence Association Xi'an Branch, main seat of IEEE GRSS Xi'an Branch IEEE TGRS Deputy Editor-in-Chief, Chief expert of the Innovation team of the Ministry of Education, IEEE Fellow, IET Fellow, the first batch of members of the Chinese artificial Intelligence Society and CCF distinguished members have been selected on the Elsevier cited Scholars list for five years in a row. He is a member of the discipline Review Group of the academic degrees Committee of the State Council, an expert of the postdoctoral Management Committee of the Ministry of Human and Social Affairs, and a former deputy to the eighth National people's Congress. In 1991, he was approved as an expert who enjoyed the government subsidy of the State Council. in 1996, he was first selected into the national "10 million" talent project (level 1 and 2), and the first batch of "three to five talents" in Shaanxi Province. He was elected as a national model teacher, an outstanding contribution expert in Shaanxi Province and a model teacher in Shaanxi Province.

02.

Li Ziqing

Outstanding challenges in face recognition

▲ Li Ziqing

Summary of the report

Deep learning technology improves the performance of face recognition by several orders of magnitude, making it widely used. However, in practical application, the face recognition system still encounters a variety of problems, which significantly degrade its performance and fail to achieve the desired results. This report discusses the key problems that have not been well solved in face recognition, analyzes the reasons, and puts forward the direction of solution.

A brief introduction to the reporter

Li Ziqing, Professor, West Lake University, IEEE Fellow. He was a senior researcher at Research Lead of Microsoft Research Asia and State key Laboratory of pattern recognition, Institute of Automation, Chinese Academy of Sciences. He has published more than 400 papers, written 9 books, and cited 36000 times in Google Scholar. He has served as deputy editor-in-chief of IEEE TPAMI and other publications, and served as the chair, program chair, or program committee member of more than 100 international academic conferences. As an expert in face recognition and intelligent video surveillance, he has presided over a number of national scientific research projects and major application projects, and has been granted and applied for more than 20 patents in related fields. EyeCU, Bill, a face recognition system developed by Microsoft. Gates was interviewed by CNN to explain it. Responsible for a number of national projects and international cooperative scientific research projects, developed face recognition system and intelligent video surveillance system, implemented and played a role in a number of major national security projects. Li Ziqing is vice chairman of SAC/TC100/SC2; on behalf of the Chinese national body, he wrote and adopted China's first ISO/SC37 international standard for biometric identification, and delivered a keynote speech on "Biometrics in China" (biometric recognition in China) at the plenary meeting of the annual meeting.

03.

Wang Gang

Autopilot doesn't have a free lunch.

▲ Wang Gang

Summary of the report

With the development of autopilot today, it is still faced with many difficulties. This sharing will discuss how to break down autopilot into simpler problems, and introduce how Alibaba and Autonomous driving lab's platform can effectively solve these problems.

A brief introduction to the reporter

Wang Gang, Director Alibaba Autonomous driving lab, Chief Scientist. Prior to that, Wang Gang was a tenured professor at Nanyang technological University in Singapore. In 2016 and 2017, he was selected as the Asian and Global TR35 Award by the MIT Technical Review Magazine, respectively. He is an expert of the National Thousand talents Plan, the editorial board of IEEE TPAMI, the top journal of artificial intelligence, and the chief chair of top meetings such as CVPR and ICCV.

04.

Feng Jiashi

Progress and Prospect of Human posture estimation Technology

▲ Feng Jiashi

Summary of the report

Human posture estimation has important applications in many fields, including intelligent security, autopilot, human-computer interaction and entertainment. This report will introduce the latest development of human posture estimation based on deep learning, including single, multi-person and 3D human pose estimation. In addition, this report will focus on the latest human posture estimation methods and models for solving practical application challenges, such as complex scenes and limited computing resources, and will explore the application of unsupervised learning in large-scale human posture estimation.

A brief introduction to the reporter

Feng Jiashi is an assistant professor in the Department of Electronic and computer Engineering at the National University of Singapore and head of the Machine Learning and Vision Laboratory at the National University of Singapore. Bachelor, Department of Automation, University of Science and Technology of China, Ph. D., Department of Electronic and computer Engineering, National University of Singapore. Engaged in postdoctoral research in the artificial Intelligence Laboratory of the University of California, Berkeley from 2014 to 2015. His research interests are image recognition, deep learning and robust machine learning for big data. He has won ICCV 2015 TASK-CV Best Paper Award, 2012 ACM Multimedia Conference Best Technical demonstration Award, served as the chair of ICMR 2017 Technical Committee, JMLR, IEEE TPAMI, IJCAI and other international well-known journals, conference reviewers, and has published more than 60 papers in the field of computer vision and machine learning.

05.

Lu Jiwen

Deep reinforcement Learning and Visual content understanding

▲ Lu Jiwen

Summary of the report

Deep reinforcement learning is a research hotspot in the field of artificial intelligence and is considered to be one of the important ways for human beings to move towards general artificial intelligence. By combining the perception ability of deep learning with the decision-making ability of reinforcement learning, deep reinforcement learning realizes the perception and decision-making from original input to semantic output in an end-to-end way, and has made an important breakthrough in many visual content understanding tasks. The report will share several deep reinforcement learning methods for visual content understanding proposed by the Intelligent Vision Lab of the Department of Automation of Tsinghua University in recent years, mainly including multi-agent deep reinforcement learning, graph deep reinforcement learning, and structured deep reinforcement learning, as well as their applications in many visual content understanding tasks, such as object detection and recognition, target tracking and retrieval, behavior prediction and recognition.

A brief introduction to the reporter

Lu Jiwen, associate professor and doctoral supervisor of Automation Department of Tsinghua University, mainly studies computer vision, machine learning and intelligent robots. He has published more than 70 papers in IEEE, more than 50 papers in CVPR / ICCV / ECCV, and been cited more than 8200 times. Presided over more than 10 scientific research projects such as the Joint key Fund of the National Natural Science Foundation and the National key R & D Program. Served as editor-in-chief of international journals PR Letters, editorial board of T-IP, T-CSVT, T-BIOM, PR and JVCI, principal chair of program committee of international conferences ICME 2020, AVSS 2020 and DICTA 2019, principal chair in the field of CVPR 2020, ICIP 2017 / 2018 / 2019, ICME 2015 / 2017 / 2018 / 2019, ICPR 2018and WACV 20182020. He was selected into the Youth Thousand talents Plan of the Organization Department of the CPC Central Committee in 2015 and the National Outstanding Youth Fund project in 2018.

06.

Lao Shixiang

Computer Vision Technology in China and Japan: the industrialization process from face recognition to Autopilot

▲ Lao Shifu

Summary of the report

A major breakthrough in computer vision technology is the industrial application of face detection. Japan's camera industry is the first to apply face detection technology, which removes an obstacle for the development of face recognition technology in the future. The development of face recognition technology has experienced half a century, until the emergence of deep learning to achieve a breakthrough in industrial application. Another major application of computer vision is autopilot. The emergence of deep learning makes us see the possibility of using computer vision to realize autopilot. The speaker will share more than 20 years of experience in technical exchanges and cooperation between China and Japan, the industrial application of face detection and face recognition, and the application of computer vision in the field of self-driving.

A brief introduction to the reporter

Lao Shicheng, Vice President of Shangtang Technology, General Manager of Intelligent driving Business, General Manager of Shangtang Japan. Responsible for the self-driving business of Shangtang Technology and the business of Shangtang Technology in Japan. Former head of face technology at OMRON in Japan. While working at OMRON, his team and Tsinghua University jointly developed the world's first commercial face detection chip known as "OKAO Vision", which was adopted by major camera and mobile phone manufacturers, and developed embedded face recognition technology, which was adopted by famous mobile phone manufacturers as the world's first smartphone with face verification function. Developed the world's first vending machine that uses gender and age inference technology to automatically recommend drinks to customers; first used face recognition technology to achieve beauty and whitening function, and was adopted by major printer manufacturers; developed the world's first driver status recognition chip and system (DMS). In 2009, it won the most authoritative SSII Takagi Award in the field of image processing in Japan.

07.

Shen Shengmei

Intelligent video image helps intelligent security construction

▲ Shenshengmei

Summary of the report

It is predicted that by 2020, there will be more than 1 billion video surveillance cameras in the world. How to store and watch a large number of video streams and quickly understand massive video content? After the camera was digitized that year, the video image compression (H.264, H.265) reduced the transmission volume and compressed the storage space, but with the emergence of high definition, the requirement of clear images, the sharp increase in the number of cameras and the rapid expansion of video stream data, it brings new challenges. Is artificial intelligence expected to compress video content, provide a means to automatically and quickly understand video content, and even achieve pre-warning, mid-event disposal, and post-analysis? Today, the intelligence of video images based on computer vision technology is developing in this direction. This report will take the solution of Pensi Technology as an example to illustrate the application of video image intelligence in the field of intelligent security.

A brief introduction to the reporter

Shen Shengmei is the chief scientist of Pengsi Science and Technology and the director of Singapore Research Institute. Master of Electronic Information Engineering, Department of Electronic Engineering, Xi'an University of Electronic Science and Technology. Shen Shengmei is a leader in the field of artificial intelligence and deep learning, former vice president of the Singapore (Panasonic) Research Institute, leading an algorithm research team of more than 40 people and accumulating more than 300 patents. It has the full stack technical ability of computer vision, and the technical area spans many application fields. the world's top achievements have been achieved in the fields of face detection and recognition based on deep learning, pedestrian detection and tracking, pedestrian re-recognition, vehicle recognition, autopilot, driver behavior detection, mobile operation robot and so on. In March 2019, he announced to join the domestic artificial intelligence company Pensi Technology as chief scientist and director of the Singapore Research Institute, dedicated to monitoring and safety, smart city, self-driving, intelligent robot and AI factory automation and other related technology research.

Round table Panel

The Future of computer Vision Technology in the Post-Deep Learning era

Moderator: Chen Xilin

Distinguished guests: Jiao Li Cheng, Li Ziqing, Wang Gang, Feng Jiashi, Lu Jiwen, Lao Shifu, Shen Shengmei

Round table introduction

On the topic of "Post-Deep Learning era, the Future of computer Vision Technology", the invited guests will discuss the academic research achievements, the current situation of technology landing, pain points and difficulties, and share the guests' views and opinions on the future trend from the perspectives of face recognition, video structure, video analysis, intelligent security, self-driving and other industries.

Https://www.leiphone.com/news/201910/nqtSIozqzPY6J3DF.html

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.