Invited Overview Session

Session TP1-1
Thursday, December 17, 14:00 - 15:30, Room Y301

Recent Advances in Energy Harvesting Communications
M.L. Willy Ku, Yan Chen, and K.J. Ray Liu, University of Maryland, College Park, USA

Abstract: Energy harvesting from ambient energy sources can potentially reduce the dependence on the supply of grid or battery energy, providing many attractive benefits to the environment and deployment. However, unlike the conventional stable energy, the intermittent and random nature of the renewable energy makes it challenging in the realization of energy harvesting transmission schemes. Extensive research studies have been carried out in recent years to address this inherent challenge from several aspects. In this talk, we present an overview of the recent developments in energy harvesting communications.

K. J. Ray Liu was named a Distinguished Scholar-Teacher of University of Maryland, College Park, in 2007, where he is Christine Kim Eminent Professor of Information Technology. He leads the Maryland Signals and Information Group conducting research encompassing broad areas of information and communications technology with recent focus on future wireless technologies, network science, and information forensics and security.

Dr. Liu was a recipient of the 2016 IEEE Leon K. Kirchmayer Technical Field Award on graduate teaching and mentoring, IEEE Signal Processing Society 2014 Society Award, and IEEE Signal Processing Society 2009 Technical Achievement Award. Recognized by Thomson Reuters as a Highly Cited Researcher, he is a Fellow of IEEE and AAAS.

Dr. Liu is a BoG membre of APSIPA and a Director-Elect of IEEE Board of Director. He was President of IEEE Signal Processing Society, where he has served as Vice President - Publications and Board of Governor. He has also served as the Editor-in-Chief of IEEE Signal Processing Magazine.

He also received teaching and research recognitions from University of Maryland including university-level Invention of the Year Award; and college-level Poole and Kent Senior Faculty Teaching Award, Outstanding Faculty Research Award, and Outstanding Faculty Service Award, all from A. James Clark School of Engineering.

Perceptual Coding: Hype or Hope?
C.-C. Jay Kuo, University of Southern California, USA

Abstract: There has been a significant progress in image/video coding in the last 50 years, and many visual coding standards have been established, including JPEG, MPEG-1, MPEG-2, H.264/AVC and H.265, in the last three decades. The visual coding research field has reached a mature stage, and the question "is there anything left for image/video coding?" arises in recent years. One emerging R&D topic is "perceptual coding". That is, we may leverage the characteristics of the human visual system (HVS) to achieve a higher coding gain. For example, we may change the traditional quality/distortion measure (i.e., PSNR/MSE) to a new perceptual quality/distortion measure and take visual saliency and spatial-temporal masking effects into account. Recent developments in this area will be reviewed. Then, one may ask furthermore "Is it sufficient to keep visual coding research vibrant and prosperous for another decade with such a modification?" It is probably not. On the other hand, we may formulate the coding problem dramatically differently from the past - making it fundamentally different with an HVS centric approach. The notion of Just-Noticeable-Differences (JND) will be introduced in this context, and numerous new R&D opportunities will arise accordingly.

Dr. C.-C. Jay Kuo received his Ph.D. degree from the Massachusetts Institute of Technology in 1987. He is now with the University of Southern California (USC) as Director of the Media Communications Laboratory and Dean's Professor in Electrical Engineering-Systems. His research interests are in the areas of digital media processing, compression, communication and networking technologies. Dr. Kuo was the Editor-in-Chief for the IEEE Trans. on Information Forensics and Security in 2012-2014. He was the Editor-in-Chief for the Journal of Visual Communication and Image Representation in 1997-2011, and served as Editor for 10 other international journals. Dr. Kuo received the National Science Foundation Young Investigator Award (NYI) and Presidential Faculty Fellow (PFF) Award in 1992 and 1993, respectively. He was an IEEE Signal Processing Society Distinguished Lecturer in 2006, and the recipient of the Electronic Imaging Scientist of the Year Award in 2010 and the holder of the 2010-2011 Fulbright-Nokia Distinguished Chair in Information and Communications Technologies. Dr. Kuo is a Fellow of AAAS, IEEE and SPIE. Dr. Kuo has guided 126 students to their Ph.D. degrees and supervised 23 postdoctoral research fellows. He is a co-author of about 230 journal papers, 870 conference papers and 13 books.

Learning Approach on Image Interpolation and Super-resolution
Wan-Chi Siu, The Hong Kong Polytechnic University, Hong Kong

Abstract: Image interpolation and super-resolution are important topics in image and video signal processing. Their applications include ultra-HDTV, image coding, image resizing, image manipulation, face recognition and surveillance. The objective is to increase the resolution of an image/video through up-sampling, deblurring, and/or denoising. However the definitions of interpolation and super-resolution are very confusing, even among researchers. In this talk we start to clarify, as fast as possible, the definitions of interpolation and super-resolution. This is followed by a highlight of our most recent learning approach for image interpolation and super-resolution. This is done via random forest and tree structures, which is the fastest approach with the quality comparable to or even better than those obtained from deep-learning methods; hence represents one of the state-of-the-art approaches on image/video interpolation and super-resolution.

Wan-Chi Siu, PhD DIC, FIEEE, received the PhD degree from Imperial College, London, in 1984, and is Fellow of the IEEE. He joined Hong Kong Polytechnic University as a Lecturer in 1980 and has been Chair Professor since 1992. He was Head of Department (EIE) and subsequently Dean of Engineering Faculty between 1994 and 2002. Professor Siu is an expert in digital signal processing, fast algorithms, video coding, 3D videos, pattern recognition and visual surveillance. He has published over 490 research papers, and has 8 recent patents. Prof. Siu was an independent non-executive director of a listed video surveillance company in Hong Kong for over 15 years. His works are well received by peers with high citations, and have been ported into hi-tech industrial uses.
Prof. Siu was a Vice President of the IEEE Signal Processing (SP) Society, Chairman of Conference Board and a core member of the Board of Governors (2012-2014). He initiated (together with other BoG members) and implemented successfully new conference series for the IEEE SP Society, and set up criteria and typical procedures for quality conference management. Recently, he has also been elected as the President-Elect (2015-2016) of the Asia-Pacific Signal and Information Processing Association (APSIPA). Prof. Siu is/was subject editor, guest Editor and associate editor of a number of IEEE and other journals, such as Electronics Letters, IEEE Transactions on Circuits & Systems for Video Technology, IEEE Transactions on Image Processing, and IEEE Transactions on Circuits and Systems. He is a very popular lecturing staff member within the University, while outside the University he has been a keynote speaker of over 12 international/national conferences in the recent 10 years. He received many awards, such as Distinguished Presenter Award, the Best Teacher Award, the Best Paper Award and IEEE Third Millennium Medal. He took up the leading role in organizing over 20 international conferences in Hong Kong, mainland China and overseas in these 30 years with high commendation, including say for example the prestigious conferences MMSP'2008 in Australia as a Co-General Chair, ICIP'2010 as General Chair, ICASSP'2003 as General Chair and ISCAS'1997 as TPC Chair, where the last three are IEEE Society-sponsored flagship international conferences. In 1992/3, he chaired the First Engineering/IT Panel of the Research Assessment Exercise (RAE) and initiated to set up a set of objective indicators to assess the basic research quality of academia, which gives substantial impact to the research culture in Hong Kong for the recent 22 years.

Session FP1-1
Friday, December 18, 14:00 - 15:30, Room Y301

Virtual-View Based 3D Video Composition
Hsueh-Ming HANG, National Chiao Tung University, Taiwan

Abstract: One interesting next-generation 3D research direction is the so-called virtual-viewpoint (or free-viewpoint) video system. It is also an on-going standardization item in the international ITU/MPEG Standards. Typically, a densely arranged camera array is used to acquire input images and a number of virtual view pictures are synthesized at the receiver using the depth-image based rendering (DIBR) technique. An interesting application of virtual-view system is 3D scene composition. It is an extension of the traditional chroma key technique but it now tries to merge two sets of 3D video scenes into one consistent 3D scene. However, these two sets of RGB-D sequences are taken independently by two different sets of cameras. Thus, the camera orientations and movements of these cameras may not match each other. We will discuss the challenges of this topic and summarize our progress on solving them.


Hsueh-Ming Hang received the B.S. and M.S. degrees from National Chiao Tung University, Hsinchu, Taiwan, in 1978 and 1980, respectively, and Ph.D. in Electrical Engineering from Rensselaer Polytechnic Institute, Troy, NY, in 1984. From 1984 to 1991, he was with AT&T Bell Laboratories, Holmdel, NJ, and then he joined the Electronics Engineering Department of National Chiao Tung University (NCTU), Hsinchu, Taiwan, in December 1991. From 2006 to 2009, he took a leave from NCTU and was appointed as Dean of the EECS College at National Taipei University of Technology (NTUT). He is currently the Dean of the ECE College, NCTU. He has been actively involved in the international MPEG standards since 1984 and his current research interests include multimedia compression, image/signal processing algorithms and architectures, and multimedia communication systems.

Dr. Hang holds 13 patents (Taiwan, US and Japan) and has published over 190 technical papers related to image compression, signal processing, and video codec architecture. He was an associate editor (AE) of the IEEE Transactions on Image Processing (1992-1994, 2008-2012) and the IEEE Transactions on Circuits and Systems for Video Technology (1997-1999). He is a co-editor and contributor of the Handbook of Visual Communications published by Academic Press in 1995. He was a Board Member of the Asia-Pacific Signal and Information Processing Association (APSIPA) (2013-2014) and currently an IEEE Circuits and Systems Society Distinguished Lecturer (2014-2015). He is a recipient of the IEEE Third Millennium Medal and is a Fellow of IEEE and IET and a member of Sigma Xi.

hide details

From Algorithm/Architecture Co-Exploration to Internet of Things
Gwo Giun (Chris) Lee, National Cheng Kung University, Taiwan

Abstract: NIKLAUS EMIL WIRTH introduced the innovative idea that Programming = Algorithm + Data Structure. Inspired by this, the current talk advances the idea to the next level by stating that Design = Algorithm + Architecture. With concurrent exploration of both algorithm and architecture entitled Algorithm/Architecture Co-exploration (AAC), this methodology introduces a leading paradigm shift in advanced system design including cloud computing and Internet of Things.

As computing becomes exceedingly demanding and data becomes increasingly bigger, efficient parallel and flexible reconfigurable processing are crucial in the design of signal processing systems. Hence the analysis of algorithms for potential computing in parallel is crucial. AAC presents a technique which systematically lays out the full spectrum of potential parallel processing components eigen-decomposed into possible data granularities. With data dependency minimization, this spectrum of independent graph components is resolved from a particular data granularity into lower and mixed granularities within the design space. This makes possible the study of potentials for homogeneous or heterogeneous parallelization at different granularities as opposed to conventional systolic array for homogeneous designs at single fixed granularity with possible extensions to distributed computing on cloud platforms. Because AAC was targeted for SoC systems with versatile platforms, the scope of system is extensible to systems connected via signals in conveying information thus forming Internet of Things (IoT). This introduces a fundamental framework for general system design already with major impact to SoC systems and has been broaden to cloud computing, Deep Learning in machine learning, and even genomic and proteomic signal processing systems in bioinformatics.

Gwo Giun (Chris) Lee (S'91-M'97-SM'07) received his B.S. degree in Electrical Engineering from National Taiwan University and both his M.S. and Ph.D. degrees in Electrical Engineering from University of Massachusetts. Dr. Lee has held several technical and managerial positions in the industry including System Architect in former Philips Semiconductors, USA, DSP Architect in Micrel Semiconductors, USA, and Director of Quanta Research Institute, Taiwan before joining the faculty team of the Department of Electrical Engineering in National Cheng Kung University (NCKU) in Tainan, Taiwan where he established the Media SoC Laboratory. He was also a visiting Professor at "Swiss Federal Institute of Technology" (EPFL), Switzerland during 2007. Dr. Lee has authored more than 200 technical papers and is currently a member of the ISO/IEC MPEG standardization committee and was also the chief editor for the Reconfigurable Video Coding (RVC) Ad Hoc group. He was the Chair for the Complexity Analysis Ad HoC Group of ISO/IEC ITU JVT-3V in 3D Video Coding. Dr. Lee also serves as the Associate Editor for both IEEE Transactions on Circuits and Systems for Video Technology from 2009 till 2013 and Journal of Signal Processing Systems since 2010.
He received the Best Associate Editor's Award for IEEE Transactions on Circuits and Systems for Video Technology in 2010 and the Best Paper Award for the BioCAS track in ISCAS 2012. Dr. Lee was also the Guest Editor for IEEE TCSVT's November, 2009 special issue on "Algorithm/ Architecture Co-Exploration for Visual Computing on Emergent Platforms". He is the Chair of the technical committee for "Visual Signal Processing & Communications" track and member of "Multimedia Systems Application" track for IEEE International Symposium on Circuits and Systems (ISCAS). Dr. Lee also serves as the technical committee member for both the Digital Implementation of Signal Processing Systems (DISPS) and the Industry Digital Signal Processing (IDSP) committees for IEEE Signal Processing Society and Circuits and Systems Society. Furthermore, he is currently the Chair of the Signal Processing Systems Track in Asia Pacific Signal and Information Processing Association (APSIPA). His research interests are focused on intelligent and biomedical algorithm, architecture, VLSI/SoC design, and Algorithm/Architecture Co-Exploration (AAC) for signal and information processing systems including cloud computing and Internet-of-Things.

A MIMO-OFDM System for High-Quality Video Communication
Yoshikazu Miyanaga, Hokkaido University, Japan

Abstract: Currently sophisticated wireless technologies have enabled high-speed data transmission in home and personal networks. As a wireless communication standard, IEEE802.11ac based wireless LAN supports the maximum throughput of 1.5 Gbps at a 40-MHz frequency band width (BW), 3.0 Gbps at 80-MHz BW and 6.0 Gbps at 160-MHz BW by using a multiple-input and multiple-out (MIMO) stream technique with orthogonal frequency division multiplexing (OFDM).
This high throughput can be in particular expected for the use of high quality video wireless communications. In this talk, new wireless systems "over 1G" bps throughput, "over 80MHz" bandwidth and "less than 6GHz" carrier are introduced. In addition, some results in the field experiments are introduced when such high speed wireless systems are applied for high quality video transmission.
Our developed system has achieved the data rate of 3 Gbps by use of an 80-MHz baseband bandwidth and a 8 x 8 MIMO scheme. This talk describes the VLSI implementation of the 8 x 8 MIMO-OFDM system. A low-latency and the optimum pipelined architecture are employed for all processing blocks to provide the real-time operations on OFDM modulation and MIMO detection. The proposed architecture also realizes low power consumption. This system has been applied for High-Quality Video communication. With some of results on field experiments, the system performance for video communications is described.

Yoshikazu Miyanaga is the dean and a professor of Graduate School of Information Science and Technology, Hokkaido University. His research interests are in the areas of speech signal processing, wireless communications and low-power consumption VLSI design. He is an associate editor of Journal of Signal Processing, RISP Japan (2005-present). He was President-elect, IEICE Engineering Science (ES) Society (2014-2015) and currently President (2015-present). He is a fellow member of IEICE. He was a vice-President (2009-2013), Asia-Pacific Signal and Information Processing Association (APSIPA) and now a member of the APSIPA advisory committee. He was a distinguished lecture (DL) of IEEE CAS Society (2010-2011), an associate editor of IEEE CAS Transaction on TCAS-II (2011-2013) and he was a Board of Governor (BoG) of IEEE CAS Society (2011-2013).

hide details

Session SA2-1
Saturday, December 19, 11:00 - 12:30, Room Y301

Transfer Learning for Speech and Language Processing
Dong Wang and Thomas Fang Zheng, Tsinghua University, China

Abstract: Transfer learning is a vital technique that generalizes models trained for one setting or task to other settings or tasks. For example in speech recognition, an acoustic model trained for one language can be used to recognize speech in another language, with little or no re-training data. Transfer learning is closely related to multi-task learning (cross-lingual vs. multilingual), and is traditionally studied in the name of 'model adaptation'. Recent advance in deep learning shows that transfer learning becomes much easier and more effective with high-level abstract features learned by deep models, and the 'transfer' can be conducted not only between data distributions and data types, but also between model structures (e.g., shallow net and deep net) or even model types (e.g., Bayesian model and neural model). This review paper summarizes some recent prominent research towards this direction, particularly for speech and language processing. We also report some results from our group and highlight the potential of this very interesting research field.

Dr. Dong Wang received the B.Sc. and M.Sc. degrees in computer science from Tsinghua University in 1999 and 2002. He then joined Oracle China during 2002-2004 and IBM China during 2004-2006. He joined CSTR, University of Edinburgh, in 2006 as a Marie Curie Research Fellow, where he received the PhD degree in 2010. From 2010 to 2011, he was with EURECOM as a Postdoctoral Fellow, and from 2011 to 2012, was a Senior Research Scientist with Nuance. He is now an Assistant Professor with Tsinghua University, Beijing, China. Dr. Wang works on speech processing, language processing and finance signal processing. The collaborative work with several commercial partners leads to the open Lingyun AI cloud service used by millions of people every day.

Signal and Information Processing Applications for the Smart Grid
Anthony Kuh, University of Hawaii, USA

Abstract: This talk discusses using signal processing to assist in processing of information for the smart grid. This consists of getting information about the electrical grid and environment via sensor networks, interpreting information received via signal processing and machine learning, and then using the information to make intelligent decisions about the grid using control and optimization algorithms. The focus is on the electrical grid beyond the last substation, the distribution grid. For the smart distribution grid there is an increasing amount of distributed renewable energy sources and possible distributed storage. This necessitates gathering more information about the electrical grid, environment data, and building energy usage. With this information we can forecast distributed renewable energy sources and develop algorithms for distributed state information. We can then develop demand response algorithms to control loads (e.g. appliances, thermostats, air conditioners, hot water heaters). While this talk is an overview talk we discuss some details of our research efforts in these areas.

Anthony Kuh received his B.S. in Electrical Engineering and Computer Science at the University of California, Berkeley in 1979, an M.S. in Electrical Engineering from Stanford University in 1980, and a Ph.D. in Electrical Engineering from Princeton University in 1987. Dr. Kuh previously worked at AT&T Bell Laboratories and has been on the faculty in Electrical Engineering at the University of Hawai'i since 1986. He is currently a Professor in the Department and is also currently serving as director of the interdisciplinary renewable energy and island sustainability (REIS) group. Previously, he served as Department Chair of Electrical Engineering Dr. Kuh's research is in the area of neural networks and machine learning, adaptive signal processing, sensor networks, communication networks, and renewable energy and smart grid applications.

Dr. Kuh won a National Science Foundation Presidential Young Investigator Award and is an IEEE Fellow. He was also a recipient of the Boeing A. D. Welliver Fellowship and received a Distinguished Fulbright Scholar's Award working at Imperial College in London. Dr. Kuh was an Associate Editor for the IEEE Transactions on Circuits and Systems, served on the IEEE Neural Networks Administrative Committee, served on the IEEE Neural Networks for Signal Processing Committee, and was a Distinguished Lecturer for the IEEE Circuits and Systems Society. Dr. Kuh co-chaired the 1993 International Symposium on Nonlinear Theory and Its Applications (NOLTA) and served as the technical co-chair for the 2007 IEEE ICASSP both held in Honolulu. He was serving as the IEEE Signal Processing Society Regions 1-6 Director at Large (2013-2014). He is currently on the Board of Governors of the Asia Pacific Signal and Information Processing Association, and as a senior editor of the IEEE Journal of Selected Topics in Signal Processing.

hide details

Session SP1-1
Saturday, December 19, 14:00 - 15:30, Room Y301

Sparse Adaptive Filters -an Overview and Some Emerging Trends
Mrityunjoy Chakraborty, Indian Institute of Technology, Kharagpur, India

Abstract: In practice, one often encounters systems that have a sparse impulse response (IR), with the degree of sparseness varying over time. Examples of such systems include network echo channels in voice and data communication, wireless multipath channels in mobile communication, echo channels in HDTV, acoustic channels in underwater communication etc. The a priori information about sparseness of the system IR, if exploited properly, can significantly improve the identification performance of the algorithm deployed to identify it. In recent years, several sparse adaptive filters have been proposed that cleverly incorporate the a priori knowledge about sparseness of the system in the coefficient adaptation relations and thus perform better. The first and foremost in this category is the proportionate normalized LMS (PNLMS) algorithm and its variants like the improved PNLMS (IPNLMS) and the ?-law PNLMS (MPNLMS). In the PNLMS category of algorithms, the step size for each coefficient is made proportional to the magnitude of the corresponding coefficient update, thereby making it large for active taps (leading to faster rate of convergence initially) and small for inactive taps (leading to lesser steady state excess mean square error (EMSE)). Apart from the PNLMS family, another powerful class of sparse adaptive filters has come up in recent years, inspired by the recent advent of compressive sensing in general and LASSO in particular. The primary development in this is the so-called zero attracting LMS (ZA-LMS) algorithm, obtained by adding a norm penalty (of the filter coefficient vector) to the LMS cost function. Minimization of the cost function introduces certain zero attractors in the weight update formula which pull the coefficient updates towards zero. The ZA-LMS was later modified to reweighted zero attracting LMS (RZA-LMS) where the shrinkage is restricted only to the inactive taps. The ZA-LMS and the RZA-LMS algorithms offer lesser steady state EMSE as compared to the PNLMS family while enjoying a convergence rate that is reasonably good though not as high as that of the PNLMS. In addition to the PNLMS family and the ZA-based algorithms, there have been several other approaches also to realize a sparse adaptive filter, notably, the partial update LMS, convex combination of adaptive filters etc. Further, sparse adaptive filters have been used as nodes in a distributed network deployed to identify the unknown sparse system and diffusion strategies have been devised for sharing of information within the neighborhood of each node,, resulting in refined estimates.

The purpose of this talk is to present the basics of some of the major recent developments in the context of sparse adaptive filters. No background knowledge in this area will be assumed though some familiarity with basic adaptive filtering will be helpful. It is expected that participants will gain some useful input from this talk, enabling them to pursue further studies in this area in future.

Mrityunjoy Chakraborty obtained Bachelor of Engg. from Jadavpur university, Calcutta, Master of Technology from IIT, Kanpur and Ph.D. from IIT, Delhi. He joined IIT, Kharagpur as a faculty member in 1994, where he currently holds the position of a professor in Electronics and Electrical Communication Engg. The teaching and research interests of Prof. Chakraborty are in Digital and Adaptive Signal Processing, VLSI Signal Processing, Linear Algebra and Compressive Sensing. In these areas, Prof. Chakraborty has supervised several graduate theses, carried out independent research and has several well cited publications.

Prof. Chakraborty has been an Associate Editor of the IEEE Transactions on Circuits and Systems, part I (2004-2007, 2010-2012) and part II (2008-2009), apart from being an elected member (also currently the chair elect) of the DSP Technical Committee (TC) of the IEEE Circuits and Systems Society, a guest editor of the EURASIP JASP (special issue), track co-chair (DSP track) of ISCAS 2015 & 2016, Gabor track chair of DSP-15, and a TPC member of ISCAS (2011-2014), ICC (2007-2011) and Globecom (2008-2011). Prof. Chakraborty is a co-founder of the Asia Pacific Signal and Information Processing Association (APSIPA), is currently a member of the APSIPA BOG and also, served as the chair of the APSIPA TC on Signal and Information Processing Theory and Methods (SIPTM). He has also been the general chair and also the TPC chair of the National Conference on Communications - 2012.

Prof. Chakraborty is a fellow of the Indian National Academy of Engineering (INAE) and also a fellow of the IETE. During 2012-2013, he was selected as a distinguished lecturer of the APSIPA.

hide details

Biomedical signal processing and systems' state of the arts and future research challenges
Tomasz M. Rutkowski, University of Tsukuba, Japan

Abstract: The lecture will summarize the state research and current activities of the Biomedical Signal Processing and Systems (BioSiPS) Technical Committee member labs from Asia-Paci?c. The talk is addressed to biomedical and general signal processing audience, since biomedical monitoring studies cover a wide range of signal analysis and machine learning related problems. Majority of recent hot BioSiPS applications are related to the brain data processing and online interfacing (brain-computer or brain-to-brain interfaces, etc.). There is also a growing interest in sleep studies, which are based on a fusion of biomedical signal processing methods comprising brain (EEG) and body peripheral electrophysiological (EOG, EMG, EKG, etc.), acoustic (breath and snoring sounds), body movements and temperature, skin conductance, etc. The multimodality of the above signals recorded at different scales creates new challenges for BioSiPS applications. There is also a growing interest in biomedical wearables for which energy ef?cient data processing and storage using Internet-of-Things (IoT) technologies will be also reviewed. The lecture will conclude with an outline of the future BioSiPS research challenges.

Tomasz M. RUTKOWSKI received his M.Sc. in Electronics and Ph.D. in Telecommunications and Acoustics from Wroclaw University of Technology, Poland, in 1994 and 2002, respectively. He received a postdoctoral training at the Multimedia Laboratory, Kyoto University, and in 2005-2010 he worked as a research scientist at RIKEN Brain Science Institute, Japan. Currently he serves as an assistant professor at the University of Tsukuba and as a visiting scientist at RIKEN Brain Science Institute. Professor Rutkowski's research interests include computational neuroscience, especially brain-computer interfacing technologies, computational modeling of brain processes, neurobiological signal and information processing, multimedia interfaces and interactive technology design. He received The Annual BCI Research Award 2014 for the best brain-computer interface project. He is a senior member of IEEE, a member of the Society for Neuroscience, and the Asia-Paci?c Signal and Information Processing Association (APSIPA) where he serves as BioSiPS Technical Committee Chairman. He is a member of the Editorial Board of Frontiers in Fractal Physiology and serves as a reviewer for "IEEE TNNLS, IEEE TSMC - Part B, Cognitive Neurodynamics, and the Journal of Neural Engineering, PLOS One, Nature Scientific Reports, etc.

hide details

Face Recognition from Low-resolution to High-resolution
Kenneth K.M. Lam, The Hong Kong Polytechnic University

Abstract: A lot of research on face recognition has been conducted over the past two decades or more. Various face recognition methods have been proposed, but investigations are still underway to tackle different problems and challenges for face recognition. The existing algorithms can only solve some of the problems, and their performances degrade in real-world applications. In this talk, we will first discuss the performances of face recognition techniques on face images at different resolution. To perform face recognition, image features from a query image are first extracted and then matched to those features in a gallery set. The amount of information and the effectiveness of the features used will determine the recognition performance. To improve the performance, we will present a face recognition approach using information about face images at higher and lower resolutions, which can enhance the information content of the features that are extracted and combined at different resolutions. As the features from different resolutions should closely correlate with each other, we introduce the cascaded generalized canonical correlation analysis (GCCA) to fuse the information to form a single feature vector for face recognition. To further improve the performance and efficiency, "Gabor-Feature Hallucination" is used to predict the high-resolution (HR) Gabor features from the Gabor features of a face image directly by using local linear regression. We also describe how the algorithm is extended for low-resolution (LR) face recognition.

For recognition of HR face images, we will show that pore-scale facial features can be explored when the resolution of faces is greater than 700x600 pixels. We will describe the use of the facial features for recognition under conditions of different facial expressions, lighting, poses and captured times. We will also present the minimum area in face images that can retain a high recognition level. Experiment results indicate that the facial pores can be used as a new biometric for recognition, even distinguishing between identical twins.

Kin-Man Lam received the Associateship in Electronic Engineering with distinction from The Hong Kong Polytechnic University (formerly called Hong Kong Polytechnic) in 1986, the M.Sc. degree in communication engineering from the Department of Electrical Engineering, Imperial College of Science, Technology and Medicine, London, U.K., in 1987, and the Ph.D. degree from the Department of Electrical Engineering, University of Sydney, Sydney, Australia, in August 1996.

From 1990 to 1993, he was a Lecturer at the Department of Electronic Engineering, The Hong Kong Polytechnic University. He joined the same department as an Assistant Professor in October 1996, became an Associate Professor in 1999, and has been a Professor since 2010. He has been a member of the organizing committee and program committee of many international conferences. Dr. Lam was also the Chairman of the IEEE Hong Kong Chapter of Signal Processing between 2006 and 2008. Between 2009 and 2013, he was an Associate Editor of IEEE Trans. on Image Processing.

Currently, Dr. Lam is VP-Member Relations and Development of the Asia-Pacific Signal and Information Processing Association (APSIPA), and the Director-Membership Services of the IEEE Signal Processing Society. He serves as an Associate Editor of Digital Signal Processing, APSIPA Trans. on Signal and Information Processing, and EURASIP International Journal on Image and Video Processing. He is also an Editor of HKIE Transactions, and an Area Editor of IEEE Signal Processing Magazine. He is a General Co-Chair of the 2015 APSIPA Annual Summit and Conference and the 2017 IEEE International Conference on Multimedia Expo, both to be held in Hong Kong. His current research interests include human face recognition, image and video processing, and computer vision.

hide details