Abstract:
Nearest neighbour computation, a critical step in diverse problems like classification, regression, retrieval etc., is indeed a major challenge while dealing with large data. Practically, however, it suffices to find the computationally simpler "approximate near neighbor" instead. This is specially true of image processing applications, where the relationships between images can be computed very efficiently with such approximations. Of late, several algorithms to compute the "approximate nearest neighbour", such as PatchMatch, CSH, FeatureMatch etc. have emerged, leading to their extensive use in a wide range of applications that include image editing, completion, retargeting, de-noising, dense correspondence fields, object detection, label propagation etc. This talk briefly introduces the above listed algorithms and their applications.
Brief Bio:
R. Venkatesh Babu received his Ph.D from the Dept. of Electrical Engg., Indian Institute of Science, Bangalore, India in 2003. He held postdoctoral positions at NTNU, Norway and IRISA/INRIA, Rennes, France, through ERCIM fellowship. Subsequently he worked as a research fellow at NTU, Singapore. He spent couple of years working in industry. Currently he is working as Assistant Professor at Supercomputer Education and Research Centre (SERC), Indian Institute of Science, Bangalore, India. His research interests include video analytics, human computer interaction, computer vision, compressed domain video processing. He is a senior member of IEEE.
Abstract:
The electromagnetic spectrum, traditionally used for wireless communication, has become an extremely scarce natural resource. While sub-10 GHz frequencies have been widely used for cellular telephony, broadcasting and wireless local area networks, higher frequencies have been off-limits due to limitations in fabricating high volume CMOS integrated circuits. Due to continued CMOS scaling, such limitations no longer exist, and consequently, the industry is slowly transitioning to operating frequencies much higher than 10 GHz. At the same time, for sub-10 GHz frequencies, new techniques are being explored to improve spectrum usage.
In this talk, I will discuss two of the research areas that are being proposed to improve the efficiency of spectrum usage "cognitive radios and 60-GHz radios. I will talk about some of the existing challenges in the design of integrated circuits for such applications.
Brief Bio:
Gaurab Banerjee received the B.Tech (Hons.). degree in Electronics and Electrical Communication Engineering from the Indian Institute of Technology, Kharagpur, India, and the Ph.D. degree in Electrical Engineering from the University of Washington, Seattle, in 1997 and 2006, respectively.
In 1999, he joined Intel Corporation, Hillsboro, OR, to design analog and mixed-signal circuits for the first Pentium-4 microprocessor. Between 2001 and 2007, he was a research scientist with Intel Labs, working on CMOS based analog, mixed-signal and RF circuits for wireless and wire-line communication systems. Between 2007 and 2010, he was a staff engineer with Qualcomm Inc., Austin, TX, working on RFIC design for mobile broadcast video applications. Since May 2010, he has been an Assistant Professor in the Department of Electrical Communication Engineering, Indian Institute of Science, Bangalore, India. His research interests are in analog and RF integrated circuits and systems for communication and sensor applications. He has published more than 20 papers on semiconductor devices and circuits and has about 10 patents granted or pending. Between 2008 and 2010, Dr. Banerjee was an Associate Editor of IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I. He has also served as a reviewer for many IEEE journals and on the technical program committees of many conferences. Dr. Banerjee is a National Talent Search Scholar of India, and a Senior Member of IEEE.
Abstract:
ESD is a serious reliability threat to semiconductor chips, however chip technology, process, device and circuit co-design requirements for ESD protection is often not known to researchers exploring device, materials, process and design options for future nanoelectronic products. This talk will provide an introduction to the essential concepts of on-chip ESD protection devices and circuits. Moreover, research options for researchers in the area of materials, nanoelectronic devices, circuits and compact modelling will be highlighted. Finally, an outlook on the ESD device research in the advanced CMOS technologies will be given.
Brief Bio:
Prof. Mayank Shrivastava has a wide experience and interest in the field of Nanoscale device design and modelling, ESD devices and circuits, device-circuit co-design, drain extended MOS devices and electrothermal modelling. He has taken several positions within the semiconductor industry. During 2008 and again in 2010, he was a Visiting Scholar at Infineon Technologies AG, Munich, Germany. During 2010-2011, he worked for Infineon Technologies, East Fishkill, NY, USA and later Intel, Mobile & Communications Group, Hopewell Junction, NY, USA. From Oct. 2011 till Aug. 2013 he was with Intel, Mobile & Communications Group, Munich Germany. Since September 2013 he is with DESE at IISc Bangalore.
Prof. Shrivastava has over 35 publications in international journals/conferences and has 18 United States patents issued or pending in his field of interest. He was a recipient of the India TR35 award for the year 2010 (Young Innovator Award from MIT Technology Review 35); 2008 Best Research Paper Award in circuit design category from Intel Corporation Asia Academic Forum; the 2010 Industrial Impact Award from IIT Bombay; the biography publication by the International Biographical Centre, Cambridge, U.K., in the 2000 Outstanding Intellectuals of the 21st Century in 2010; the Excellence in Thesis work for his Ph.D. thesis from IIT Bombay and Infineon PhD fellowship for 3 years.
Abstract:
Due to the increase in the installation of surveillance cameras, the problem of recognizing persons from their facial images captured by these cameras have become very important. Facial images captured by surveillance cameras usually have poor resolution in addition to uncontrolled poses and illumination conditions which adversely affect performance of face matching algorithms. In this talk, we will discuss these different challenges and different approaches that have been proposed to address them. This talk will also cover new and interesting areas of research in this field.
Brief Bio:
Dr. Soma Biswas is an Assistant Professor at Indian Institute of Science, Bangalore. She received her PhD from University of Maryland, College Park in 2009 and worked as a Research Assistant Professor in the Department of Computer Science and Engineering at the University of Notre Dame. She also worked as a Research Scientist at GE Global Research, Bangalore before joining IISc. Her research interests are in Computer Vision, Pattern Recognition, Image Processing and related areas.
Abstract:
We discuss convergence of the following discrete-time non-linear dynamical system: n agents are located in R^d and at every time step, each moves synchronously to the average location of all agents within a unit distance of it. This popularly studied system was introduced by Krause to model the dynamics of opinion formation and is often referred to as the Hegselmann-Krause model. We prove the first polynomial time bound for the convergence of this system in arbitrary dimensions.
Brief Bio:
Arnab Bhattacharyya is an assistant professor at the Indian Institute of Science. He received his Ph.D. in 2011 from the Massachusetts Institute of Technology. He also did postdoctoral work at Princeton University and Rutgers University. His research interests are algorithms, complexity and computational questions about natural systems.