DICTA 2015 Conference
Keynote Speakers
The DICTA 2015 committee welcomes international keynote speakers Prof. Bjarne K. Ersbøll (DTU), Dr. Francois Chaumette (INRIA), Prof. Manik Varma (Microsoft Research India) and Prof. Vincent Lepetit (TU Graz).Professor Bjarne Kjær Ersbøll, Technical University of Denmark [homepage]
Imaging Techniques and Image Analysis for Food Quality and Safety
A significant amount of applied research at DTU concerns food quality and safety. A number of different DTU departments are involved in one way or another. For the presenters department: Department of Applied Mathematics and Computer Science food quality and safety poses excellent challenges for the different disciplines represented. In this talk we will look into some different ways of data collection from food samples and food processes. Special attention will be given to a range of imaging techniques which have proven useful and also to the subsequent analysis and interpretation of such images. Some of the imaging techniques which will be presented include: multispectral imaging with a highly reproducible calibrated system – now commercialized and sold under the trade name Videometer; new X-ray imaging techniques involving phase contrast and dark field imaging; and subsurface laser scattering. Each of these techniques imply different challenges to image analysis and pattern recognition algorithms. Also, we are starting to scratch the surface into the era of Big Data with the monitoring of continuous food production and optimization processes. To help circumvent these challenges we will be looking at ways to reduce the amount of information created in intelligent and robust ways – methods include feature generation and sparse ways of analyzing the data.
|
BiographyBjarne Kjær Ersbøll’s work is mainly on applied statistics and data analysis. He has considerable experience in the application of these disciplines in industrial, medical projects. His research and teaching is largely inspired by finding solutions to actual problems in industry and other institutions - and often in collaboration with these. Bjarne Kjær Ersbøll is a Full Professor in statistics and data analysis at DTU Compute. In 1983 he received the M.Sc.(eng.) degree and in 1990 the Ph.D. degree, both from the Technical University of Denmark (DTU). Bjarne Kjær Ersbøll has been employed at DTU since 1983. First he was Research Assistant (1983-1986), then be became Assistant Professor (1986-1992), Associate Professor (1992-2008), Professor (2008-2009), Full Professor (2010-present). He gives research based consultancy in applied statistics and data analysis. He has supervised a very large number of master’s thesis projects and also Ph.D. thesis projects. Furthermore, he has organised or co-organised a large number of conferences on image analysis and statistics. |
Doctor Francois Chaumette, INRIA, France [homepage]
Visual Servoing With and Without Image Processing
Visual servoing consists in using the data provided by a vision sensor for controlling the motions of a dynamic system, robots typically. The classical approach is based on an image processing-then-control loop, extracting and tracking from the image a set of geometrical features used as input of a closed-loop control scheme. A more recent approach avoids this feature extraction step. The talk will present the main basic modeling and control aspects of these two approaches, as well as their respective advantages and drawbacks. An overview of the visual servoing applications in robotics will be also given.
|
BiographyFrançois Chaumette, IEEE RAS Fellow, is an INRIA senior research scientist at IRISA in Rennes, France, where he leads the Lagadic group since 2004. He received the M.Sc. (eng.) degree from "Ecole Nationale Supérieure de Mécanique", Nantes, in 1987 and a Ph.D. in computer science from the University of Rennes in 1990. His research interests lie in the area of robot vision, mainly visual servoing and active perception. He has published over 230 journal or conference papers, with the 2002 Best IEEE Transactions on Robotics and Automation Paper Award. He has served on the technical program committee of the main conferences in computer vision (ECCV, CVPR, ICCV) and robotics (ICRA, IROS, RSS). He has been Associate Editor of the IEEE Transactions on Robotics (2001-2005) and is currently in the Editorial Board of the Int. Journal of Robotics Research and Senior Editor of the new IEEE Robotics and Automation Letters. |
Professor Manik Varma, Microsoft Research India [homepage]
Extreme Classification: A New Paradigm for Ranking and Recommendation
In this talk, I will introduce the new and rapidly growing area of extreme classification which deals with learning problems with millions of labels/categories. Extreme classification is an important research area since it enables the tackling of web scale applications with many labels such as recognizing millions of people in images, millions of words in photo OCR, millions of objects in videos, etc. Furthermore, extreme classification has also opened up a new paradigm for ranking and recommendation with applications such as ranking images on Bing, recommending videos on YouTube, recommending hashtags on Twitter and generating image captions on Facebook.
After introducing the area, I will develop the FastXML algorithm for extreme multi-label learning. FastXML learns a hierarchy over millions of labels so as to make predictions in logarithmic time. The label hierarchy is learnt by optimizing a ranking loss function, known as nDCG, via a new alternating minimization technique. It is demonstrated that FastXML's predictions can be significantly more accurate as compared to the state-of-the-art and that FastXML can be upto a thousand times faster to train.
I will conclude by reviewing open research questions in the area and discussing potential solutions to some of them.
|
BiographyManik Varma is a researcher at Microsoft Research India. He received a bachelor's degree in Physics from St. Stephen's College, University of Delhi in 1997 and another one in Computation from the University of Oxford in 2000 on a Rhodes Scholarship. He then stayed on at Oxford on a University Scholarship and obtained a DPhil in Engineering in 2004. Before joining Microsoft Research, he was a Post-Doctoral Fellow at the Mathematical Sciences Research Institute Berkeley. He has been an Adjunct Professor at the Indian Institute of Technology (IIT) Delhi in the Computer Science and Engineering Department since 2009 and jointly in the School of Information Technology since 2011. His research interests lie in the areas of computer vision, machine learning and computational advertising. He has served as an Area Chair for machine learning and computer vision conferences such as ACCV, CVPR, ICCV, ICML, ICVGIP and NIPS. He has been awarded the Microsoft Gold Star award and has won the PASCAL VOC Object Detection Challenge. Classifiers that he has developed are running live on millions of machines around the world protecting them from viruses and malware. |
Professor Vincent Lepetit, TU Graz, Austria [homepage]
Soft Computer Vision Methods for Hard Computer Vision Problems
Early works in 3D camera and object registration, which I will call "Hard Computer Vision", were mostly based on handcrafted methods. By contrast, problems such as object detection of category recognition, or "Soft Computer Vision", are typically solved using Machine Learning and statistical methods. For the last decade, I have been revisiting "Hard Computer Vision" problems with "Soft Computer Vision" techniques, and I will present applications to keypoint detection and description, and object and hand 3D registration.
|
Biography
Vincent Lepetit is a Professor at the Institute for Computer Graphics and Vision, TU Graz and a Visiting Professor at the Computer Vision Laboratory, EPFL. He received the PhD degree in Computer Vision in 2001 from the University of Nancy, France, after working in the ISA INRIA team. He then joined the Virtual Reality Lab at EPFL as a post-doctoral fellow and became a founding member of the Computer Vision Laboratory. He became a Professor at TU GRAZ in February 2014. His research interests include vision-based Augmented Reality, 3D camera tracking, Machine Learning, object recognition, and 3D reconstruction. He often serves as program committee member and area chair of major vision conferences (CVPR, ICCV, ECCV, ACCV, BMVC). He is an editor for the International Journal of Computer Vision (IJCV) and the Computer Vision and Image Understanding (CVIU) journal. |
