iciicii2020

The Fourth International Conference on Industrial Informatics - Computing Technology, Intelligent Technology, Industrial Information Integration

Shantou, China

December 18-19, 2020

ICIICII 2020

Invited Speech

Local landscape

Co-sponsors

Zhijun Zhang, PhD，Professor

School of Automation Science and Engineering,

South China University of Technology

Title: Vary Parameter Recurrent Neural Network Applied to

Intelligent Robots and Data Analysis

Abstract

Everything in nature is eternal and absolute with time changing, while the static state is relative. Inspired by this fundamental law of nature and based on neural dynamics method, Dr. Zhijun Zhang designed and proposed a varying-parameter recurrent neural network. He designed and deduced various forms of the networks, and proved that the network has the property of super-exponential convergence in solving time-varying problems and robot motion planning problems. This model is more effective in suppressing noise when solving noisy problems and has many advantages. For instance, the network model can effectively overcome the limitations of existing methods in solving time-varying, nonlinear, underdetermined, and multi-solution problems in complex robotic systems, which can only be solved for specific types of robots with slow convergence and weak robustness, and has the advantages of high solution accuracy, fast error convergence, and strong robustness. This method can be applied to data analysis and mining, robot motion planning, natural human-robot interaction, flight controller design and many other aspects in real systems.

Biography

Zhijun Zhang received the Ph.D. degree from Sun Yat-sen University, Guangzhou, China, in 2012. He was a Post-Doctoral Research Fellow with the Institute for Media Innovation, Nanyang Technological University, Singapore, from 2013 to 2015. From 2015 to 2019, he worked as an Associate Professor, and since 2020 he has been a Full Professor with the School of Automation Science and Engineering, South China University of Technology. His current research interests include neural networks, robotics, machine learning, human machine interaction, and optimal control.

Zhijun Zhang has obtained Guangdong Science Fund for Distinguished Young Scholars, and Youth Talents in Science and Technology Innovation of Guangdong Special Support Plan. He is currently a member of Committee on Adaptive Dynamic Programming and Reinforcement Learning, a member of Committee on Visual Cognition and Computation of Chinese Society of Image Graphics, and a member of Committee on Hybrid Intelligence of Chinese Association of Automation, and a Member of Committee on Coexisting-Cooperative-Cognitive Robots. He is currently the Executive Editor-in-Chief of Global Journal of Neuroscience, Associate Editor of International Journal of Robotics and Control, Review Editor of Frontiers in Robotics and AI, reviewers of more than 20 international SCI journals, the reviewer for the National Foundation of China, and the reviewer for high-tech enterprises in Guangdong Province. He is the pioneer in variable parameter neurodynamics and neurodynamics-based method for real-time and natural human-robot interaction.

He has published more than 80 papers in international journals and conferences, and more than 20 first/corresponding IEEE Transaction papers, 2 highly cited papers, and 2 books/chapters. He received 1 Best Paper Award of International Conference IEEE ICAL2011, 1 Nomination Award of SCIS-CCC Poster Award of Science China - China Control Conference, and 1 Workshop Best Poster Award of International Robotics Flagship Conference IROS2019.

The bird's-eye view of Shantou University
Nan’ao Island
Nan'ao Bridge
Zhongshan Pavilion
Shantou Coastal Corridor
Lotus Pond
Gentleman Sculpture Group

Alex Noel Joseph Raj, PhD，Professor

Department of Electronic Engineering, College of Engineering,

Shantou University, China

Title: Hybrid U-NET models for lesion segmentation in Medical Images.

Abstract

The localization and segmentation of the lesions in Medical image images is helpful for clinical diagnosis of the disease. The talk foucses on architectures on RDAU-NET (Residual-Dilated-Attention-Gate-UNet) model, RDAU_NET-WGAN and ADID-UNET that are employed to segment lesions. The model is based on the conventional U-Net, but the plain neural units are replaced with residual units oe dense networks to enhance the edge information and overcome the network performance degradation problem associated with deep networks. To increase the receptive field and acquire more characteristic information, dilated convolutions were used to process the feature maps obtained from the encoder stages. The traditional cropping and copying between the encoder-decoder pipelines were replaced by the Attention Gate modules which enhanced the learning capabilities through suppression of background information. The proposed models, are tested with BUS images and COVID 19 lung CT images and their segmentation results as compared to other Deep Networks. The experimental results illustrate that the proposed Hybrid-UNET models can accurately segment lesions when compared to other deep learning models and thus has a good prospect for clinical diagnosis.

Biography

ALEX NOEL JOSEPH RAJ received his Ph.D. degree in Engineering from the University of Warwick, Coventry, U.K. in 2009, the Master's degree in Applied Electronics from Anna University in 2005, and the Bachelor's degree in Electrical Engineering from Madras University, India, in 2001. From October 2009 to September 2011, he was with Valeport Ltd Totnes, U.K. as Design Engineer. From March 2013 to Dec 2016 he was with the Department of Embedded Technology, School of Electronics Engineering, VIT University, Vellore, India as a Professor. Since January 2017, he is with Department of Electronic Engineering, College of Engineering, Shantou University, China. He is specialized in Image processing, with Industrial and Teaching experience in Machine Learning, Deep Networks, Signal and Medical Image Processing, and FPGA implementations.

Last

First