Prof. Yao Zhao is a distinguished Changjiang Scholar, a Distinguished Young Scholar of NSFC, a leader in scientific and technological innovation of the Ten Thousand Talents Program, IEEE Fellow. He is currently the director of the Institute of Information Science at Beijing Jiaotong University and the director of the Beijing Key Laboratory of Modern Information Science and Network Technology. His research field is digital media information processing and intelligent analysis, including image/video compression, digital media content security, media content analysis and understanding, artificial intelligence, etc. He is leading or led over 30 projects including the 2030 New Generation Artificial Intelligence Project, the 973 Plan, and the 863 Plan for technological innovation. He has published over 200 papers in international journals and conferences, including IEEE Transactions. As the first prize winner, Prof. Zhao has won 5 provincial and ministerial level awards such as the first prize of the Beijing Science and Technology Award. Eight doctoral students under his guidance won the Excellent Doctoral Dissertation Award of Beijing and China Computer Federation. He was invited to serve as an editorial board member for multiple international magazines, including IEEE Transactions on Cybernetics, IEEE Transactions on Circuits and Systems for Video Technology. He is a member of the State Council Discipline Evaluation Group for the discipline of "Information and Communication Engineering" and an expert of the Cloud Computing and Big Data Special Project of the Key R&D Program of the Ministry of Science and Technology.
Image semantic segmentation is an interdisciplinary research direction involving computer vision, pattern recognition, and artificial intelligence. It is a key scientific issue in applications such as autonomous driving, intelligent monitoring, virtual reality, medical image diagnosis, and robotics. At present, deep learning has made significant breakthroughs in the field of image semantic segmentation. However, a large number of pixel level annotations typically require a significant amount of time, money, and manpower. Therefore, the insufficient or missing training data has become one of the key factors restricting the further development of image semantic segmentation.
In order to reduce the huge burden of pixel level annotation, many weakly supervised image semantic segmentation techniques have been proposed in recent years, which utilize a large amount of easily obtainable weakly supervised information (such as image labels) to complete more complex image semantic segmentation tasks.
Interactive semantic segmentation is an important technical means to reduce the cost of pixel level annotation by guiding computers to achieve fast and accurate object segmentation through simple human-computer interaction.
This report will introduce some research works of my research group in weakly supervised image semantic segmentation and interactive image segmentation.
Prof. Nannan Wang
Xidian University, China
Wang Nannan, Huashan distinguished professor at Xidian University, is currently the associate director of State Key Laboratory of Integrated Services Networks (ISN). In recent years, he has been engaged in the research of computer vision and machine learning. His research mainly involves cross-domain image reconstruction and credible identity authentication, including sketch-photo synthesis and recognition, image/video super-resolution reconstruction, image restoration, behavior analysis and recognition, person re-identification, etc. He has published over 150 papers in top international journals and conferences such as IEEE TPAMI, IJCV, CVPR, ICCV, ECCV, NeurIPS, ICML, etc. He has received Outstanding Youth Foundation from National Natural Science Foundation of China. He has been selected as Young Elite Scientists Sponsorship Program by China Association of Science and Technology (CAST). He has been awarded the first prize for Ministry of Education Natural Science Award, the first prize for Shaanxi Province Science and Technology Award, the second prize of China Society of Image and Graphics (CSIG) Natural Science Award. He is the recipient of the Chinese Association for Artificial Intelligence (CAAI) Outstanding Doctorate Dissertations Award and Shaanxi Province Outstanding Doctorate Dissertations Award. He is the associate editor-in-chief of The Visual Computer.
Cross-domain Image Synthesis
Cross-domain images are images with different representations obtained by the same target through different sensors or acquisition means. Cross-domain synthesis refers to generating an image of another domain from an image of one domain using the correlation of content and the complementarity of expression between cross-domain images. This research topic includes many sub-fields. Face sketch-photo synthesis, image super-resolution, medical image synthesis, and image style transfer all belong to the scope of this topic. This lecture discusses the topic of cross-domain image synthesis from the following four aspects:
(1) Background, giving the definition and import applications of cross-domain image synthesis;
(2) Research progress, reviewing the development history of cross-domain image synthesis and introducing some representative studies; (3) Recent advances, exploring cutting-edge techniques and new tasks in this field;
(4) Challenges and future directions, discussing the typical challenges and highlighting the future research directions.
Assoc. Prof. Hongliang Ren
Associate Professor The Chinese University of Hong Kong (CUHK), Hong Kong Adjunct Associate Professor National University of Singapore, Singapore
Hongliang Ren received his Ph.D. in Electronic Engineering (Specialized in Biomedical Engineering) from The Chinese University of Hong Kong (CUHK) in 2008. He serves as an Associate Editor for IEEE Transactions on Automation Science & Engineering (T-ASE) and Medical & Biological Engineering & Computing (MBEC). He has navigated his academic journey through Chinese University of Hong Kong, Johns Hopkins University, Children’s Hospital Boston, Harvard Medical School, Children’s National Medical Center, United States, and National University of Singapore (NUS). He is currently Associate Professor, Department of Electronic Engineering at Chinese University of Hong Kong, and Adjunct Associate Professor, Department of Biomedical Engineering at National University of Singapore. His areas of interest include biorobotics, intelligent control, medical mechatronics, soft continuum robots, soft sensors, and multisensory learning in medical robotics. He is the recipient of NUS Young Investigator Award and Engineering Young Researcher Award, IAMBE Early Career Award 2018, Interstellar Early Career Investigator Award 2018, ICBHI Young Investigator Award 2019, and Health Longevity Catalyst Award 2022 by NAM & RGC.
Surgical motion generation and motion understanding towards augmented minimally invasive robotic procedures
This talk will highlight some recent developments in dexterous robotic motion generation with motion understanding towards image-guided minimally invasive procedures. The procedure-specific telerobotic surgical systems can assist surgeons in performing dexterous manipulations using the master-slave console bilateral motion generation & mapping mechanism with variable stiffness. Meanwhile, surgical motion understanding aims to learn from the multi-domain surgical perceptions and describe the semantic relationship between instruments and surgical region of interest. Automatically understanding the instrument motions in robotic surgery is crucial to enhance surgical outcomes, enable surgical camera automation, and facilitate surgical training. To that end, we generate the task-aware saliency maps and scanpath of the instruments beyond tracking and segmentation, similar to the surgeon’s visual perception, to get the priority focus on selected surgical instruments. Furthermore, generating a surgical report in robot-assisted surgery, together with surgical scene understanding, can play a significant role in document entry tasks, surgical training, and post-operative analysis.
Prof. Xiuling Liu
Xiuling Liu, is a professor and doctoral supervisor. She is the recipient of Hebei Outstanding Youth Fund, the second level candidate of Hebei "3 3 3 Talent Project", one hundred excellent innovative talents in Hebei Province, and the vice president of Hebei Machine Learning Society. From September 1995 to February 2002, she graduated from the School of Electrical and Automation Engineering of Tianjin University with a bachelor's degree and a master's degree, and then worked as a system analyst at Trans.cosmos China, and was introduced to the School of Electronic Information Engineering of Hebei University at the end of 2004. Now she is the Dean of School of Electronic Information Engineering of Hebei University. She is mainly engaged in medical system simulation, imaging system modeling and signal analysis research. His main research in the past five years includes: intelligent analysis and 3D visualization of multimodal cardiovascular images, physical modeling and surgical simulation of human soft tissues, virtual simulation of coronary stenosis evaluation from morphological to functional transformation, feature analysis and feature wave extraction of ECG signals, intelligent analysis and diagnosis of cardiovascular diseases based on ECG signals.
Prof. Yoshihiro Yamanishi
Yoshihiro Yamanishi mainly focuses on Bioinformatics, Drug discovery, Inference, Drug side effects and Drug. His Bioinformatics study typically links adjacent topics like Computational biology. His Computational biology study combines topics in areas such as Biological data, Genome, Drug target, KEGG and In silico. He interconnects Diagnostic marker, Disease gene, Kernel canonical correlation analysis and Kegg pathway in the investigation of issues within Genome. His work in Inference covers topics such as Phylogenetic tree which are related to areas like Data mining, Chromosome and Inference engine. In his study, Drug withdrawal, Intensive care medicine and Drug reaction is inextricably linked to Drug development, which falls within the broad field of Drug side effects. Yoshihiro Yamanishi mostly deals with Computational biology, Artificial intelligence, Data mining, Bioinformatics and Drug discovery. His study in Computational biology is interdisciplinary in nature, drawing from both KEGG, Gene, Interactome and Drug. His Artificial intelligence research incorporates elements of Machine learning, Protein–protein interaction and Pattern recognition. His Data mining research is multidisciplinary, incorporating elements of Quantitative structure–activity relationship, Interpretability, Support vector machine and Metabolic pathway. Yoshihiro Yamanishi combines subjects such as Interaction network, Diagnostic marker, G protein-coupled receptor and Genomics with his study of Bioinformatics. The Drug discovery study combines topics in areas such as Plasma protein binding, Data-driven, Transcriptome, In silico and Binding site.