Plenary Speaker

/uploads/image/2022/03/15/图片1.png

Prof. Wen-Feng Wang

International Academy of Visual Art and Engineering, London, UK

Prof. Wen-Feng Wang is an evaluation expert for National Natural Science Foundation of China, National Excellent Youth Fund (overseas projects), National Publishing Fund and Shanghai Government Procurement, and Shanghai Enterprise Technological Innovation Project. He is the Editor in Chief of International Journal of Applied Nonlinear Science, one editorial board member of Nature-scientific reports, the general chair of the 3DWCAI, the chief scientist of Shanghai Lingang Artificial Intelligence Lab, the chief scientist in the field of big data and intelligent computing of RealMax, and in 2021, he was selected as "100 people in the intelligent era" of China. He is now a professor of Shanghai Institute of Technology, a tenured professor of IMT Institute in India and the director of Sino-Indian Joint Research Center of Artificial Intelligence and Robotics. He has been invited as a reviewer for tens of SCI journals, including some top ones - Nature Computational Science, Expert System with Applications, Water Research, Science of the Total Environment, Environmental Pollution, IEEE Transactions on Automation Science and Engineering. He has been invited as keynote speakers of many influential Springer-Nature conferences and has been invited to give special reports to many famous institutes in China and U.S. He motivated the foundation of the International Academy of Visual Arts and Engineering in London in 2021 to promote the novel applications of artificial intelligence in visual arts and visual engineering. As a well-known scholar in artificial intelligence, he has participated in THE's Global Academic Reputation Survey at the invitation of THE to determine the World University Reputation Ranking in 2022 and the World University Ranking in 2023.

Title:Extenics in face recognition: from a complex 2D scene to the 3D scene

Abstract:

This speech aims to explore applications of the five intelligence layers in face recognition access control manufacturing for tackling the great challenge for occluded face recognition in 3D visualization scene, which embedded the occluded faces problem into a 3D visualization scene to develop extenics in face recognition. By fusing the video containing faces in a complex 2D scene with the 3D model of the same scene, the faces images in 3D visualization scene are obtained and recognized. Theory of extenics in face recognition is essentially necessary for fighting against convid-19, when occluded face recognition technology can serve as access control in public places with a large number people to solve the contradiction between face recognition and wearing masks. Extension of applications in face recognition access control manufacturing also include establishing the experimental platform for validation of extenics in face recognition and developing human-machine interfaces. The involved algorithms and models are also explained.


/uploads/image/2022/02/16/图片2.png

Prof.Xudong Jiang

Nanyang Technological University, Singapore

Xudong Jiang (Fellow of IEEE) received the B.Eng. and M.Eng. from the University of Electronic Science and Technology of China (UESTC), and the Ph.D. degree from Helmut Schmidt University, Hamburg, Germany. From 1998 to 2004, he was with the 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. Currently, he is a Assoc. Professor in School of EEE, NTU. Dr Jiang  holds 7 patents and has authored over 200 papers with over 40 papers in the IEEE journals, where 30 papers in IEEE Signal Processing Society Journals and 6 papers in IEEE T-PAMI. Three of his journal papers have been listed as the top 1% highly cited papers in the academic field of Engineering by Essential Science Indicators. He also published 38 papers in IEEE-SP conferences ICASSP/ICIP/ICME and 13 papers in top conferences CVPR/ICCV/ECCV. He served as IEEE-SPS IFS TC Member from 2015 to 2017, Associate Editor for IEEE SPL from 2014 to 2018 and Associate Editor for IEEE T-IP from 2016 to 2020. Dr Jiang is currently an IEEE Fellow and serves as Senior Area Editor for IEEE T-IP and Editor-in-Chief for IET Biometrics. His current research interests include Signal/image processing, machine learning, pattern recognition and computer vision.

Speech Title:Semantic Segmentation: Dense Prediction by Deep Learning

Abstract:Scene segmentation is a challenging task as it need classify every pixel in the image, a dense prediction problem. It is crucial to exploit context information and aggregate multi-scale features to achieve better dense prediction. Context is essential for dense prediction. Due to the diverse shapes of objects and their complex layout in various scene images, the spatial scales and shapes of contexts for different objects have very large variation. It is thus ineffective or inefficient to aggregate various context information from a predefined fixed region. In this talk, I will first present a novel context contrasted local feature that not only leverages the informative context but also spotlights the local information in contrast to the context. Furthermore, I will present a scheme of gated sum to selectively aggregate multi-scale features for each spatial position. The gates in this scheme control the information flow of different scale features. Finally, I will present a scale- and shape-variant semantic mask for each pixel to confine its contextual region. Using the inferred spatial scope of the contextual region, a shape-variant convolution is controlled by the shape mask that varies with the appearance of input. In addition, this work also proposes a labeling denoising model to reduce wrong predictions caused by the noisy low-level features.

Keywords:Machine Learning, Computer Vision, Artificial Intelligence, Deep Learning


/uploads/image/2022/01/24/图片1.png

Prof.Matthias F. Carlsohn

ENGINEERING & CONSULTANCY DR. CARLSOHN, Germany

Dr. Carlsohn has a track record in scientific and commercial image processing and communication with 40+ years of experiences in this field gathered from affiliations in large consumer and in defense industry, as co-founder of related start-ups as well as from several universities in Germany, France and Austria where he was appointed. In his Engineering and Consultancy for Computer Vision and Image Communication he is dealing with related commercial and industrial developments and applications and has driven the past 16 years Springer’s Journal of Real-Time Image Processing as co-editor-in-chief.

Speech Title:Real-time processing aspects in an:“Eco-system for 4D-particle image velocimetry in transparent 3D-prints of morphological synthetic resin models of cerebral aneurysms as digital twins”

Abstract:The eco-system is basing on an approved experimental setup for corpuscular flow measurement (PIV: particle image velocimetry) on transparent 3D printed synthetic resin models as patient-specific, morphological replicas of their often randomly diagnosed cerebral aneurysms – the digital twin. Its data are obtained from 3D DSA imaging with subsequent image segmentation of the vascular system supplying the aneurysm. Flowed through with a physiologically pulsating fluid of blood-like viscosity and enriched with polymer particles in the size of erythrocytes, visual inspection with a light field camera allows high-resolution particle flow analysis, represented as color-coded combination of 3D position measurement and 3D speed vectors according to spatial direction and speed (4D). These models can are used in stroke units for training and education purposes and to prepare interventions in a simplified experimental setup without a light field camera in order to practice the intervention in vitro. The models are used both for a vivid visualization in patient counselling and for doctors to practice the procedure with a catheter on a simplified experimental setup, and to prepare for special anatomical features and alternatives. In addition, the appropriate size of the stent can be estimated in advance during training on the model in order to rule out multiple attempts. Flow films and images help the doctor to decide whether an intervention is sensible and necessary at all. The risk for patients is reduced through better preparation for the intervention as well as the justified waiver of such a procedure. The costs for comparable silicone models are reduced by a factor > 1,000; their production is shortened from weeks to few hours! The eco-system consists of the measurement set-up and the logistics, receiving radiological standard data, convert them for 3D printing, generating flow images and videos with and without inserted devices, and returning models and videos via express.

Keywords:Real-time image processing, Computational modelling, 3D-printing, light field camera, plenoptic imaging, 4D-image analysis, particle image velocimetry, cerebral aneurysm, digital twin.


/uploads/image/2022/02/16/图片1.png

Prof.Jon Hall

Open University, UK

I am a leading research scientist in computing, requirements engineering, formal methods and problem solving with over 100 scientific publications and extensive experience through the design, management and evaluation of problem solving research programmes in many domains. I have successfully contributed value through my research and scholarship to many organisations, and academic, scientific and leadership committees and boards, and provided business evaluations for many companies.