Palm-Vein Verification Using Images From the RGB Spectrum

Prof. Kar-Ann Toh, School of Electrical and Electronic Engineering, Yonsei University, Seoul, Korea

Organized by Centre for Information Sciences and Systems (CISS), School of EEE, NTU& IEEE Circuits and Systems Singapore Chapter & IEEE Signal Processing Singapore Chapter

Date : 26 August 2021 (Thursday)
Time : 02.00 PM – 03.00 PM
Venue : Online Seminar (Zoom)

Abstract

In this work, we investigate into utilization of images from the visible light (RGB) spectrum for identity verification based on the palm-veins. This is differentiated from the commonly utilized Near-infrared (NIR) images for palm-vein feature extraction. Our goal is to explore into the often omitted palm-vein information from the RGB palm images considering the vast deployment of the RGB cameras. Essentially, the vein line features are extracted at various scales based on an efficient difference image projection. The extracted features from the gallery and the probe images are matched based on a fast Hamming distance implementation. The resultant similarity scores are finally fused at score level for accuracy enhancement. Experiments are conducted on two public multi-spectral palm databases. The results show encouraging matching accuracy and computational efficiency of the proposed method which extracts the palm-vein utilizing only the visible spectrum. The outcome of this study can be deployed as a standalone biometric or as part of a multibiometric system for secure authentication. 

Speaker Biography

Kar-Ann Toh is a Professor in the School of Electrical and Electronic Engineering at Yonsei University, South Korea. He received the PhD degree from Nanyang Technological University (NTU), Singapore in 1999. He worked for two years in the aerospace industry prior to his post-doctoral appointments at research centers in NTU from 1998 to 2002. He was affiliated with the Institute for Infocomm Research in Singapore from 2002 to 2005 prior to his current appointment in Korea. He was a Visiting Professor at National University of Singapore in the year 2020 during his sabbatical leave. His research interests include biometrics, pattern classification, and machine learning. He has served/is serving as an Associate Editor of several international journals including IEEE Transactions on Biometrics, Behavior and Identity Science, IEEE Transactions on Information Forensics and Security, Journal of Franklin Institute, Pattern Recognition Letters, and IET Biometrics.

Back to Top

Order Learning and Its Applications to Computer Vision  

Prof. Chang-Su Kim, School of Electrical Engineering, Korea University

Organized by IEEE Signal Processing Singapore Chapter & IEEE Circuits and Systems Singapore Chapter & Singapore University of Technology and Design (SUTD)

Date : 24 August 2021 (Tuesday)
Time : 02.00 PM
Venue : Online Seminar (Zoom)

Abstract

First, we propose order learning to determine the order graph of classes, representing ranks or priorities, and classify an object instance into one of the classes. To this end, we design a pairwise comparator to categorize the relationship between two instances into one of three cases: one instance is 'greater than,' 'similar to,' or 'smaller than' the other. Then, by comparing an input instance with reference instances and maximizing the consistency among the comparison results, the class of the input can be estimated reliably.  

Second, we propose the deep repulsive clustering (DRC) algorithm of ordered data for effective order learning. To this end, we develop the order-identity decomposition (ORID) network to divide the information of an object instance into an order-related feature and an identity feature. Then, we group object instances into clusters according to their identity features using a repulsive term. Moreover, we estimate the rank of a test instance, by comparing it with references within the same cluster.  

Experimental results on facial age estimation, aesthetic score regression, and historical color image classification show that the proposed algorithm can cluster ordered data effectively and also yield excellent rank estimation performance. 

Speaker Biography

Chang-Su Kim received the Ph.D. degree in electrical engineering from Seoul National University with a Distinguished Dissertation Award in 2000. From 2000 to 2001, he was a Visiting Scholar with the Signal and Image Processing Institute, University of Southern California, Los Angeles. From 2001 to 2003, he coordinated the 3D Data Compression Group in National Research Laboratory for 3D Visual Information Processing in SNU. From 2003 to 2005, he was an Assistant Professor in the Department of Information Engineering, Chinese University of Hong Kong. In Sept. 2005, he joined the School of Electrical Engineering, Korea University, where he is a Professor. His research topics include image processing, computer vision, and machine learning. He has published more than 300 journal and conference papers. In 2009, he received the IEEK/IEEE Joint Award for Young IT Engineer of the Year. In 2014, he received the Best Paper Award from Journal of Visual Communication and Image Representation (JVCI). He is a member of the Multimedia Systems & Application Technical Committee (MSATC) of the IEEE Circuits and Systems Society. Also, he was an APSIPA Distinguished Lecturer for term 2017-2018. He served as an Editorial Board Member of JVCI and an Associate Editor of IEEE Transactions on Image Processing. He is a Senior Area Editor of JVCI and an Associate Editor of IEEE Transactions on Multimedia. 

Back to Top

Random Walk on a Tree for Stochastic Optimization and Learning 

Prof. Qing Zhao, Joseph C. Ford Professor of Engineering, School of Electrical and Computer Engineering, Cornell University, Ithaca, NY 14853

Organized by IEEE Signal Processing Singapore Chapter & IEEE Circuits and Systems Singapore Chapter & Singapore University of Technology and Design (SUTD)

Date : 12 August 2021 (Thursday)
Time : 08.30 PM
Venue : Online Seminar (Zoom)

Abstract

The problem of searching for a few rare events of interest among a massive number of possibilities is ubiquitous. The rare events may represent opportunities with exceptional returns, extremely useful information in a deluge of data, or anomalies with potentially catastrophic consequences. The key challenges are that the search space is massive, observations are noisy and costly, and stochastic models of the rare events are unknown. Example applications include identifying infected individuals in a large population, detecting intrusions and attacks in large communication/computer networks, and the general problem of stochastic optimization for finding the optimal point of an unknown objective function in a high-dimensional space. We discuss in this talk a solution framework and its optimality in terms of learning efficiency. The key idea of the approach is to devise a biased random walk on a tree-based hierarchical representation of the search space. This is a joint work with Sudeep Salgia and Sattar Vakili. 

Speaker Biography

Qing Zhao joined Cornell University in 2015, where she is the Joseph C. Ford Professor of Engineering. Prior to that, she was a professor with the ECE Department at University of California, Davis. She received the Ph.D. degree in electrical engineering from Cornell University in 2001. Professor Zhao is a Fellow of IEEE, a Marie Skłodowska-Curie Fellow of the European Union research and innovation program, a Jubilee Chair Professor of Chalmers University during her 2018-2019 sabbatical leave, and a Distinguished Lecturer of the IEEE Signal Processing Society. She was the recipient of the 2010 IEEE Signal Processing Magazine Best Paper Award and the 2000 Young Author Best Paper Award from IEEE Signal Processing Society. Her research interests include sequential decision theory, stochastic optimization, machine learning, and algorithmic theory with applications in infrastructure, communications, and social-economic networks.   

Back to Top

Online Multiple Object Tracking based on Joint-Detection-and-Embedding Network

Dr. Zhenyu Weng, School of EEE, Nanyang Technological University, Singapore

Organized by Robotics & Automation Research Lab (ROAR), SUTD & IEEE Signal Processing Singapore Chapter & IEEE Circuits and Systems Singapore Chapter & Teochew Doctorate Society, Singapore

Date : 01 June 2021 (Tuesday)
Time : 03.00 PM – 04.00 PM
Venue : Online Seminar (Zoom)

Abstract

Most of online multiple object tracking methods consist of two subtasks, detection and embedding, and thus they need two different networks. To reduce the complexity, recent methods for tracking persons integrate these two subtasks into a unified network. In this talk, we cover two topics. Firstly, different from the above methods focusing on designing networks, we explore the online association strategy to better associate the tracks with the detected objects after performing detection and extracting embedding features from the detected objects. Secondly, we propose a unified network for face detection and embedding, and then use the proposed association strategy for multiple face tracking.

Speaker Biography

Zhenyu Weng received the B.S. degree in Computer Science and Technology at Sun Yat-sen University, China and the Ph.D. degree from School of Electronics Engineering and Computer Science at Peking University, China. He is currently working as a research fellow in School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore. His research interests cover machine learning, computer vision, deep learning, and incremental learning.

Back to Top

Stretching Beyond Linear Methods for Pattern Classification

Prof. Kar-Ann Toh, School of Electrical and Electronic Engineering, Yonsei University, Seoul, Korea

Organized by IEEE Systems, Man and Cybernetics Singapore Chapter & Chinese and Oriental Language Information Processing Society & Teochew Doctorate Society, Singapore & IEEE Circuits and Systems Singapore Chapter

Date : 27 January 2021 (Wednesday)
Time : 05.00 PM – 06.00 PM
Venue : Online Seminar (Zoom)

Abstract

According to Herbert A. Simon, “the more relevant patterns at your disposal, the better your decisions will be.” In this talk, an overview of key ideas in pattern classification will be illustrated right after a brief background introduction. Subsequently, several deterministic learning methods for regression and classification shall be introduced. These learning methods are found to be related to each other through data transformation. Finally, we introduce a stretchable learning where feature extraction and target fitting can be performed at the same time. Some numerical examples will be given to demonstrate the effectiveness of the learning method. The talk shall be concluded by a pictorial view of relationships among these classifiers.

Speaker Biography

Kar-Ann Toh is a Professor in the School of Electrical and Electronic Engineering at Yonsei University, South Korea. He received the PhD degree from Nanyang Technological University (NTU), Singapore in 1999. He worked for two years in the aerospace industry prior to his post-doctoral appointments at research centers in NTU from 1998 to 2002. He was affiliated with the Institute for Infocomm Research in Singapore from 2002 to 2005 prior to his current appointment in Korea. He was a Visiting Professor at National University of Singapore in the year 2020 during his sabbatical leave. His research interests include biometrics, pattern classification, and machine learning. He has served/is serving as an Associate Editor of several international journals including IEEE Transactions on Biometrics, Behavior and Identity Science, IEEE Transactions on Information Forensics and Security, Journal of Franklin Institute, Pattern Recognition Letters, and IET Biometrics.

Back to Top

An Asymmetric Kernel for Compressed Classification

Prof. Kar-Ann Toh, School of Electrical and Electronic Engineering, Yonsei University, Seoul, Korea

Organized by IEEE Signal Processing Singapore Chapter & IEEE Circuits and Systems Singapore Chapter & Centre for Information Sciences and Systems (ICSS), School of EEE, NTU & Singapore University of Social Sciences (SUSS)

Date : 07 January 2021 (Thursday)
Time : 02.30 PM – 03:30 PM
Venue : Online Seminar (Zoom)

Abstract

According to Herbert A. Simon, “the more relevant patterns at your disposal, the better your decisions will be.” In this talk, we introduce an asymmetric kernel for compressed learning representation. The kernel can be utilized for stretchable learning where feature compression and target fitting can be performed at the same time. The learning is subsequently extended to classifier learning where an error counting objective is desired. Some numerical examples on benchmark datasets will be given to demonstrate the effectiveness of the learning method.

Speaker Biography

Kar-Ann Toh is a Professor in the School of Electrical and Electronic Engineering at Yonsei University, South Korea. He received the PhD degree from Nanyang Technological University (NTU), Singapore in 1999. He worked for two years in the aerospace industry prior to his post-doctoral appointments at research centers in NTU from 1998 to 2002. He was affiliated with the Institute for Infocomm Research in Singapore from 2002 to 2005 prior to his current appointment in Korea. He was a Visiting Professor at National University of Singapore in the year 2020 during his sabbatical leave. His research interests include biometrics, pattern classification, and machine learning. He has served/is serving as an Associate Editor of several international journals including IEEE Transactions on Biometrics, Behavior and Identity Science, IEEE Transactions on Information Forensics and Security, Journal of Franklin Institute, Pattern Recognition Letters, and IET Biometrics.

Back to Top