Keynote Speakers
Prof. Guoping Qiu The University of Nottingham, UK & China Chief Scientist, Everimaging Ltd.
Speech Title: High Dynamic Range – The Last Frontier of Digital Imaging
Abstract: Many years of research and development plus billions of dollars investment in technology have made digital photography device ubiquitous and very sophisticated. Despite huge progress, there are still the occasions, for example when taking a photo of an evening party at a restaurant, where the image quality will still come out poorly. Either the dark shadows are too dark such that no details are visible, or the light areas are so bright such that they are completely saturated, and no details are visible. Even after turning on the high dynamic range (HDR) function in your camera which is now a feature in every smartphone, or adjusting the various control buttons, the situations will not improve much. And yet the photographer on the scene can clearly see every detail both in the dark and in the bright regions. The question is, why? In this talk I will show that this difficulty is caused by the high dynamic range of the light intensities of the scene, and we call this the HDR problem. I will show from first principle that HDR is the cause of many difficulties in digital imaging (photography) and correct some of the misconceptions in many recent literatures on image processing problems such as low-light (or dark) image enhancement, especially those so-called end-to-end blackbox solutions based on deep learning. I will demonstrate both theoretically and in practice that HDR is the last technical obstacle, the last frontier, of digital imaging.
Prof. Hua Huang
Beijing Normal University, China
黄华，北京师范大学人工智能学院教授，国家杰出青年基金获得者，中国计算机学会常务理事、教育工委副主任，中国图象图形学学会常委理事、多媒体专委会主任。主要研究图像与视频处理，计算摄像学。先后主持国家自然科学基金重点项目、国家重点研发计划项目等，近年来，发表IEEE Trans论文50余篇，授权国家发明专利60余项，部分成果在国防、工业、互联网等领域得到应用，获得了第十四届中国青年科技奖，入选了第三批万人计划科技创新领军人才。
Dr. Huang Hua is a professor at the School of Artificial Intelligence, Beijing Normal University. He currently serves as an executive board member and deputy director of the Academic Working Committee in the China Computer Federation (CCF), as well as an executive board member and director of the Multimedia Technical Committee in the China Society of Image and Graphics (CSIG). His research interests include image and video processing and computational photography. In recent years, he has published more than 50 papers in IEEE Transactions and has been granted more than 60 national invention patents. Some of his achievements have been applied in the fields of national defense, industry, and the Internet. He is currently presiding over a key project of the National Natural Science Foundation of China and a national key R&D plan project. He has won the 14th Science & Technology Award for Young and Middle-aged Talents and The National Science Fund for Distinguished Young Scholars.
Speech Title: TBA
Abstract: More information will be added soon...
Prof. Hui Zhang
Hunan University, China
张辉，教授，博士生导师，湖南大学机器人学院常务副院长、机器人视觉感知与控制技术国家工程研究中心副主任、中国图象图形学学会理事兼副秘书长。入选教育部“长江学者”特聘教授，国家“万人计划”青年拔尖人才，主要从事机器人视觉检测、深度学习图像识别、智能制造机器人技术及应用。
近年来主持科技创新2030—“新一代人工智能”重大项目课题、国家自然科学基金重点项目（2项），JW1XX工程重点项目，国家重点研发计划子课题、国家科技支撑计划项目子课题等20余项。在IEEE汇刊等国内外期刊上发表70多篇论文，授权国家发明专利42项，计算机软件著作权5项，获2018年国家技术发明二等奖1项，第1完成人主持获得2022年湖南省科技进步一等奖、2019年湖南省科技进步二等奖、2019年中国商业联合会科技进步奖一等奖，以主要完成人获得省部级科学技术进步奖15项，2022年湖南省第十三届教学成果特等奖，2022年高等教育（研究生）国家级教学成果奖二等奖。
Hui Zhang, Professor, Ph.D. Supervisor, serves as the Executive Vice Dean of the School of Robotics at Hunan University, Deputy Director of the National Engineering Research Center for Robot Vision Perception and Control Technology, and Council Member and Deputy Secretary-General of the China Society of Image and Graphics. He has been recognized as a Distinguished Professor under the Ministry of Education's "Changjiang Scholars Program" and a Youth Top-notch Talent of the National "Ten Thousand Talents Program." His research focuses on robotic vision inspection, deep learning-based image recognition, and intelligent manufacturing robot technologies and applications.
In recent years, he has led more than 20 projects, including a key topic under the Science and Technology Innovation 2030—"New Generation Artificial Intelligence" Major Project, two key projects funded by the National Natural Science Foundation of China, a JW1XX Engineering Key Project, sub-projects under the National Key R&D Program, and sub-projects under the National Science and Technology Support Program. He has published over 70 papers in international and domestic journals such as IEEE Transactions and holds 42 national invention patents and 5 software copyrights. He received the 2018 National Technological Invention Award (Second Prize) as the first contributor, and as the principal investigator, he was awarded the First Prize of Hunan Provincial Science and Technology Progress Award in 2022, the Second Prize in 2019, and the First Prize of the Science and Technology Progress Award of the China General Chamber of Commerce in 2019. Additionally, as a key contributor, he has won 15 provincial and ministerial science and technology progress awards, the Special Prize of the 13th Hunan Provincial Teaching Achievement Award in 2022, and the Second Prize of the National Teaching Achievement Award (Higher Education - Postgraduate) in 2022.
Speech Title: Multimodal Intelligent Perception Technology and Applications of UAVs in Complex Power Scenarios (复杂电力场景下无人机多模态智能感知技术及应用)
Abstract:
针对复杂电力场景下无人机巡检任务中的红外热故障识别、线路树障分类、杆塔倾斜检测等问题，提出了基于多模态信息融合的智能感知技术。通过结合可见光图像、红外图像、点云与多光谱数据，解决了环境复杂性、信息不完全性和传感器感知局限等挑战，显著提升了无人机系统在复杂环境中的感知与认知能力。报告重点包括：1）提出了自适应图像配准与预测信息迁移技术，解决了多模态数据的空间对齐问题，精确定位电力设备并进行温度解译；2）设计了基于点云与多光谱数据融合的树障分类方法，充分利用不同模态间的互补性，精准识别电力走廊中的树种，提升了巡检任务的精度与效率；3）开发了多模态信息协同的杆塔倾斜检测与语义分割技术，增强了复杂环境下电力设施巡检的智能化水平。通过多源数据融合与智能处理，本报告展示了如何利用多模态感知技术，提升无人机巡检在复杂电力场景中的效率、准确性和安全性，切实满足国家战略需求。
To address challenges in drone inspection tasks for complex power scenarios, including infrared thermal fault detection, line vegetation classification, and tower tilt detection, this report proposes intelligent perception technologies based on multimodal information fusion. By integrating visible light images, infrared images, point cloud data, and multispectral data, it overcomes challenges such as environmental complexity, information incompleteness, and sensor perception limitations, significantly enhancing the perception and cognition capabilities of UAV systems in complex environments. The report focuses on the following key aspects: 1) Adaptive image registration and predictive information transfer techniques are proposed to address the spatial alignment of multimodal data, enabling precise localization of power equipment and accurate temperature interpretation; 2) A tree obstacle classification method based on point cloud and multispectral data fusion is designed, leveraging the complementarity of different modalities to accurately identify tree species in power corridors, thereby improving the accuracy and efficiency of inspection tasks; 3) Multimodal information-coordinated tower tilt detection and semantic segmentation technologies are developed, enhancing the intelligence level of power facility inspection in complex environments. Through multimodal data fusion and intelligent processing, this report demonstrates how multimodal sensing technologies can improve the efficiency, accuracy, and safety of drone inspections in complex power scenarios, effectively meeting the demands of national strategic needs.
Prof. Yiu-Ming Cheung (FIEEE, FAAAS, FIET, FBCS)
Hong Kong Baptist University, Hong Kong, China
Speech Title: TBA
Abstract: More information will be added soon...
KWONG Tak Wu Sam, Lingnan University Hong Kong, China
Fellow of Hong Kong AES, US NAI and IEEE
Kwong is currently Chair Professor at the Lingnan University of the Department of Computing and Decision Science. He previously served as Department Head and Professor from 2012 to 2018 at the City University of Hong Kong. Prof Kwong joined CityU as a Department of Electronic Engineering lecturer in 1989. Prof. Kwong is the associate editor of leading IEEE transaction journals, including IEEE Transactions on Evolutionary Computation, IEEE Transactions on Industrial Informatics, and IEEE Transactions on Cybernetics. He was the President of IEEE Systems, Man And Cybernetics Society from 2022-23.
Speech Title: Deep Learning-Based Video Coding and its Applications
Abstract: In 2016, Cisco released the White paper, VNI Forecast and Methodology 2015-2020, which predicted that by 2020, 82 percent of Internet traffic would come from video applications such as video surveillance and content delivery networks. The report also revealed that in 2015, Internet video surveillance traffic nearly doubled, virtual reality traffic quadrupled, TV grew by 50 percent, and other applications similarly saw significant increases. The report estimated that the annual global traffic would first time exceed the zettabyte (ZB; 1000 exabytes [EB]) threshold in 2016 and will reach 2.3 ZB by 2020, with 1.886 ZB attributed to video data.
Today, AI and machine learning are increasingly being used in video processing to improve video quality, reduce bandwidth requirements, and enhance user experience. For instance, AI algorithms can optimize video encoding parameters based on the content of the video, reducing the bitrate required for a given level of video quality. AI can also be used for video content analysis, enabling automated scene detection, object recognition, and event detection. This has significant applications in video surveillance, where AI algorithms can be used to identify and track individuals or objects of interest in real-time.
Overall, the use of AI in video is a rapidly growing field with immense potential for improving the efficiency and quality of multimedia services. In this talk, I will present the latest research results on machine learning and deep neural network-based video coding, and their applications to the real world, such as saliency detection and underwater imaging.
Prof. Xudong Jiang, Nanyang Technological University, Singapore
IEEE Fellow
Xudong Jiang (Fellow, IEEE) received the bachelor’s and master’s degrees from University of Electronic Science and Technology of China, and the Ph.D. degree Helmut Schmidt University, Hamburg, Germany. From 1998 to 2004, he was with Institute for Infocomm Research, A*STAR, Singapore, as a Lead Scientist, and the Head of the Biometrics Laboratory. He joined Nanyang Technological University (NTU), Singapore, as a Faculty Member in 2004, where he served as the Director of the Centre for Information Security from 2005 to 2011. He is currently a professor with the School of EEE, NTU and serves as Director of Centre for Information Sciences and Systems. He has authored over 200 papers with over 60 papers in IEEE journals including 10 papers in T-PAMI and 18 papers in T-IP, and over 30 papers in top conferences such as CVPR/ICCV/ECCV/AAAI/ICLR/NeurIPS. His current research interests include computer vision, machine learning, pattern recognition, image processing, and biometrics. Dr. Jiang served as Associate Editors for IEEE SPL and IEEE T-IP. Currently he is Fellow of IEEE and serves as a Senior Area Editor for IEEE T-IP and the Editor-in-Chief for IET Biometrics.
Speech Title: How Deep CNN Revolutionizes MLP and How Transformer Revolutionizes CNN
Abstract: Discovering knowledge from data has many applications in various artificial intelligence (AI) systems. Machine learning from the data is a solution to find right information from the high dimensional data. It is thus not a surprise that learning-based approaches emerge in various AI applications. The powerfulness of machine learning was already proven 30 years ago in the boom of neural networks but its successful application to the real world is just in recent years after the deep convolutional neural networks (CNN) have been developed. This is because the machine learning alone can only solve problems in the training data but the system is designed for the unknown data outside of the training set. This gap can be bridged by regularization: human knowledge guidance or interference to the machine learning. This speech will analyze these concepts and ideas from traditional neural networks to the deep CNN and Transformer. It will answer the questions why the traditional neural networks fail to solve real world problems even after 30 years’ intensive research and development and how CNN solves the problems of the traditional neural networks and how Transformer overcomes limitation of CNN and is now very successful in solving various real world AI problems.