Computer Vision And Pattern Analysis Laboratory Home Page  Home
People  People
Publications  Publications
Publications  Databases
Contact Information  Contact
Supported Research Projects  Supported Research Projects
Research Activites  Research Activites
Research Groups
SPIS - Signal Processing and Information Systems Lab.SPIS - Signal Processing and Information Systems Lab.
Medical Vision and Analysis Group  Medical Research Activities
Biometrics Research Group  Biometrics Research Group
SPIS - Signal Processing and Information Systems Lab.MISAM - Machine Intelligence for Speech Audio and Multimedia.
Knowledge Base
  Paper Library
  VPA Lab Inventory
  Databases in VPALAB
Multimedia Recordings

Multimedia Recordings
TUM Seminars

Date and Place: April 5, 2010, 13:40-14:30 @ FENS 1040 - Sabanci University
Title: Advance Imaging and Visualization for Computer Assisted Interventions: motivation, state-of-art and future challenges
Speaker: Prof. Dr. Nassir Navab, TU Munich, Germany

Title: Medical Image Analysis for Vascular Structures: Novel Navigation Techniques for Abdominal Catheterizations
Speaker: Prof. Dr. Martin Groher, TU Munich, Germany

Title: Variational Deformable Registration Beyond L2: Overview over Approaches for Deformable Registration in Sobolev Spaces
Speaker: Darko Zikic, TU Munich, Germany

Recorded on : 2010/04/05
Filesize : ~ 2.5 GB
Length : ~ 3:30 hrs
Ground Truth Editor Demo

This program reads at input MR images and allows an operator to manually segment volumes of interest by labeling slices. GTE saves files in VTK format. The visualization is done with our 3D viewer.

For more information, please take a look at:
Ground Truth Editor
Medical Vision and Analysis Group

Recorded on : 2009/03/12
Filesize : 2MB
Length : 01:32 mn
3D Viewer Demo

This program aims at the visualization of 3D data in VTK format. One can superimpose volumetric and 3D data (surfaces).

For more information, please take a look at:
3D Viewer
Medical Vision and Analysis Group

Recorded on : 2009/03/12
Filesize : 12MB
Length : 03:27 mn
Segmentation Demo

In the framework of IronDB project, we provide results of brain Basal Ganglia organs segmentations, such as Caudate Nucleus and Putamen.

For more information, please take a look at:
Medical Vision and Analysis Group

Recorded on : 2009/03/12
Filesize : 10MB
Length : 03:34 mn
NOMA: Negatif-Olmayan Matris Ayrıştırma

Place: Sabanci University - FENS
Speaker: Bülent Sankur
"NOMA: Negatif-Olmayan Matris Ayrıştırma" Prof.Dr. Bülent Sankur, Boğaziçi Üniversitesi

Bu konuşmada, matris ayrıştırma teknikleri ve NOMA’ nın rolü, NOMA çeşitlemeleri ve uyarlamalı NOMA, NOMA’ nın nesne tanıma, biyometri, video işleme ve işaret özütleme uygulamaları üzerinde durulacak.
Biyografi: Prof. Dr. Bülent Sankur doktora derecesini Rensselaer Polytechnic Institute’ dan 1976 yılında aldı. Delft University, Ecole Nationale Superieure des Telecommunications gibi araştırma kurumlarında ziyaretçi öğretim üyesi, Marmara Araştırma Merkezi, Arçelik gibi endüstri kurumlarına danışman pozisyonlarında bulundu. Uzun süreden beri Boğaziçi Üniversitesi sinyal işleme merkezi BUSIM’ in direktörlüğünü yapmaktadır. Görüntü ve video işleme, gizli damgalama, biyometri, 3D biçim analizi, Bayes yaklaşımlı işaret işleme gibi geniş bir yelpazede ilgi alanlarına sahiptir ve bu konularda pek çok makalesi bulunmaktadır.

19 Haziran 2008 Perşembe 13:30, MDBF G032

"NMF: Non-negative Matrix Factorization" Prof. Dr. Bülent Sankur, Boğaziçi University

In this talk, the focus will be on matrix factorization techniques and NMF, types of NMF and adaptive NMF, applications of NMF in object recognition, biometry, video processing and signal extraction.
Biography: Prof. Dr. Bülent Sankur has received his Ph.d. degree at Rensselaer Polytechnic Institute at 1976. He held visiting positions at research institutes such as Delft University and Ecole Nationale Superieure des Telecommunications and consultant positions in industrial corporations such as Arçelik and Marmara Research Center. He has been the director of Boğaziçi University signal processing center BUSİM. His research interests span the areas of image and video processing, watermarking, biometry, 3D shape analysis, brain interfaces and Bayesian signal processing and he has authored or co-authored many publications on these fields.

19th June 2008 Thursday 13:30, FENS G032

Recorded on : 2008/06/19
Filesize : ~ 850 MB
Length : ~ 1 hr
Ensemble MLP Classifier Design

Place: Sabanci University - FENS
Speaker: Terry Windeatt
The idea of combining multiple classifiers is based on the observation that achieving optimal performance in combination is not necessarily consistent with obtaining the best performance for a single (base) classifier. However, the base classifier parameters still need to be set, and the optimal parameters may be different for the ensemble. The normal way to set parameters is to use a validation set or cross-validation techniques. In this talk, measures for setting parameters for two-class problems will be discussed, and extended to the problem of identifying and removing irrelevant features. The technique is extended to multi-class problems using ECOC (Error-Correcting Output Coding). Examples using MLP base classifiers for face recognition will be described.

Terry Windeatt received the B.Sc. degree in Applied Science from University of Sussex, followed by M.Sc. in Electronic Engineering from University of California and a PhD from University of Surrey, U.K. After lecturing at Kingston University, UK, he worked in the USA for eight years, at the Research and Development Departments of General Motors and Xerox Corporation in Rochester, NY. His industrial R&D experience is in modeling/simulation for intelligent automotive and office-copying applications. He returned to the UK in 1984 to the University of Surrey, where he now lectures in Machine Intelligence. He has worked on various research projects at the CVSSP, and his current research interests include neural nets, pattern recognition, and computer vision, with emphasis to facial expression analysis.

June 3, 2008, 10:40, FENS G032

Recorded on : 2008/06/03
Filesize : 636 MB
Length : 60 mn
Aytul Ercil's interview on 'TRT2' about Uykucu

Aytul Ercil's interview on 'TRT2' about Uykucu.

Recorded on : 2007/09/22
Filesize : 1.93 MB
Length : 1:07 mn
DSPinCARS Biennial Coverage on NTV

The International Alliance for Advanced Studies on In-Car Human Behavioral Signals has organized the 2007 Biennial on DSP for in-Vehicle and Mobile Systems to take place on Sunday-Tuesday, June 17-19, 2007 at the historic Sait Halim Pasa Mansion, Istanbul.

This biennial was organized to bring together scholars working on the latest techniques, standards, and emerging deployment on this central field of living at the age of wireless communications, smart vehicles, and human-machine-assisted safer and comfortable driving.

Four invited lectures, 32 papers in 3 lecture/poster sessions and a panel discussion were presented on the topics.

Recorded on : 2007/06/23
Filesize : 3.75 MB
Length : 2:28 mn
Multiple Classifier Fusion

Place: Sabanci University
Speaker: Professor Josef Kittler - Centre for Vision, Speech and Signal Processing, University of Surrey
The merits of combining multiple classifier systems are now well recognised. The talk will provide an overview of the methodology of multiple classifier fusion. The various fusion paradigms will be briefly reviewed and the underlying theory of multimodal and intramodal fusion presented. The various mechanisms of intramodal fusion will then be discussed, including Bayesian marginalisation techniques and margin manipulation techniques. Some of the resulting fusion strategies will be illustrated on selected applications.

Recorded on : 2006/04/21
Filesize : 447 MB
Length : 1h 21mn
Automatic Annotation of Sports Videos Using Contextual Reasoning

Place: SIU2006 Conference - Antalya
Speaker: Professor Josef Kittler - Centre for Vision, Speech and Signal Processing, University of Surrey
The problem of automatic annotation of sports videos is considered. It is argued that at all levels of interpretation, spatio-temporal contextual information plays a very important role. We describe a unified apparatus for reasoning in context. It is cast in the Bayesian framework of evidential reasoning. Most of the work to date has used context at a single level. However, we show that by using a multilevel context model, erroneous decisions at a lower level can be corrected by the influence of the higher levels. We illustrate, how the framework can be applied to different aspects of tennis video analysis, ranging from mosaic building for figure-ground separation, through tennis ball tracking, to the high level task of awarding points.

Recorded on : 2006/04/19
Filesize : 346 MB
Length : 1hr 3mn
Halcon Seminar

Place: SIU2006 Conference - Antalya
Speaker: Dr. Heiko Eisele - MVTec

Recorded on : 2006/04/19
Filesize : 972 MB
Length : 3 hrs
Extracting discrete information from a continuous world: Quantization, Compression, and Classification

Place: SIU2006 Conference - Antalya
Speaker: Professor Robert M. Gray - Department of Electrical Engineering, Stanford University
Scientists and engineers often seek to measure, communicate, store, process, reproduce, or analyze signals encountered in the real world. Most such signals are inherently continuous or analog in nature, yet increasingly the means for communicating, storing, and manipulating such information are discrete or digital. Generally something is lost when continuous information is converted into discrete approximations, so a natural goal is to preserve as much of the original information as possible. This is the general problem of quantization, a technique that historically has cropped up in a variety of branches of signal processing, taxonomy, physics, mathematics, and statistics as well as playing a key role as the interface between a continuous world and digital processing. Quantization traditionally has been used to model analog to digital conversion, Shannon source coding, and data compression. Viewed generally, quantization also models the extraction of information from signals, including statistical classification, clustering methods, and machine learning. This talk will describe the fundamentals of quantization along with examples and recent research topics in theory and application.

Recorded on : 2006/04/18
Filesize : 359 MB
Length : 1hr 6mn
SIU2006 Gala Yemegi - Odul Toreni (Dinner and Awards)

SIU2006 Gala Yemegi - Odul Toreni (Dinner and Awards)

Recorded on : 2006/04/18
Filesize : 345 MB
Length : 1hr 3mn
Online Handwritten Mathematical Expression Recognition

Place: SIU2006 Conference - Antalya
Speaker: Hakan Buyukbayrak
We describe a system for recognizing online, handwritten mathematical expressions, together with a user-interface for writing scientific articles. A neural network recognizes strokes and a recursive algorithm parses the expression by combining neural network output and the structure of the expression. The interface associated with the proposed system integrates the built-in recognition capabilities of the Microsoft's Tablet PC API for recognizing textual input and supports conversion of hand-drawn figures into PNG format. This enables the user to enter text, mathematics and draw figures in a single interface. After recognition, all output is combined into one LATEX code and compiled into a PDF file.

Recorded on : 2006/04/18
Filesize : 124 MB
Length : 23 mn
SPEECHFIND: Spoken Document Retrieval for a National Gallery of the Spoken Word

Place: SIU2006 Conference - Antalya
Speaker: Professor John Hansen - Department of Electrical Engineering, University of Texas at Dallas
The problem of reliable speech recognition for information retrieval is a challenging problem when data is recorded across different media and equipment. In this paper, we address the problem of audio stream phrase recognition for a new National Gallery of the Spoken Word (NGSW)[1]. This will be the first large-scale repository of its kind, consisting of speeches, news broadcasts, and recordings that are of significant historical content. An NSF initiative was recently established to provide better transition of library services to digital format. As part of this Phase-II Digital Libraries Initiative, researchers from Michigan State Univ. (MSU) and Center for Robust Speech Systems (Univ. of Texas, Dallas) have teamed to establish a fully searchable, online WWW database of spoken word collections that span the 20th Century. The database draws primarily from holdings of MSU's Vincent Voice Library which include +60,000 hours of recordings (from T.Edison's first cylinder recordings, to famous speeches such as man's first steps on the moon 'One Small Step for Man', to American presidents over the past 100 years). In this partnership, MSU will house the NGSW collection, as well as digitize (with assistance from LDC), catalog, organize, and provide meta-tagging information. MSU is also responsible for a number of engineering challenges such as digital watermarking and effective compression strategies. The team at CRSS-UTD is responsible for developing the robust automatic speech recognition and segmentation for transcript generation, and for the proto-type online audio/metadata/transcript based user search engine (

Recorded on : 2006/04/18
Filesize : 336 MB
Length : 1hr 3mn
Energy Conservation in Adaptive Filtering

Place: SIU2006 Conference - Antalya
Speaker: Professor Ali H. Sayed - Electrical Engineering Department, University of California at Los Angeles
Adaptive filters are systems that respond to variations in their environment by adapting their internal structure in order to meet certain performance specifications. Such systems are widely used in communications, biomedical applications, signal processing, and control. The performance of an adaptive filter is evaluated in terms of its transient behavior and its steady-state behavior. The former provides information about how fast a filter learns, while the latter provides information about how well a filter learns. Such performance analyses are usually challenging since adaptive filters are, by design, time-variant, nonlinear, and stochastic systems. For this reason, it has been common in the literature to study different adaptive schemes separately due to the differences that exist in their update equations.

The purpose of this talk is to provide an overview of an energy conservation approach to the performance analysis of adaptive filters. The framework is based on studying the energy flow through successive iterations of an adaptive filter and on establishing a fundamental energy conservation relation; the relation bears resemblance with Snell's Law in optics and has far reaching consequences to the study of adaptive schemes. In this way, many new and old results can be pursued uniformly across different classes of algorithms.

In particular, the talk will highlight some recently discovered phenomena pertaining to the learning ability of adaptive filters. It will be seen that adaptive filters generally learn at a rate that is better than that predicted by least-mean-squares theory; that is, they are "smarter" than originally thought! It will also be seen that adaptive filters actually have two distinct rates of convergence; they learn at a slower rate initially and at a faster rate later; perhaps in a manner that mimics the human learning process.

Recorded on : 2006/04/17
Filesize : 325 MB
Length : 1h
SIU2006 - Kurultay Acilisi (Welcome Speech)

Place: SIU2006 Conference - Antalya
Speaker: Aytul Ercil, Mujdat Cetin

Recorded on : 2006/04/17
Filesize : 77 MB
Length : 14mn
Image Indexing and Retrieval

Place: SIU2006 Conference - Antalya
Speaker: Theo Gevers - Associate Professor, Faculty of Science, University of Amsterdam
The growing capacity of computers, the abundance of digital cameras, and the increased connectivity of the world all point to large digital multimedia archives. They include images and videos from the World Wide Web, museum objects, flowers, trademarks, and views from everyday life. The faster these archives grow, the more prominent becomes the need for efficient access to the content of the images and videos.

In this short course, a survey is given of the most recent developments on image and video search engines. First, the important step of feature extraction will be discussed in detail including color, shape, and texture information, with particular attention to discriminatory power and invariance. Then the focus is on the concepts of indexing and genre classification as an intermediate step to sort the data. Attention is paid to interactive ways to perform browsing and retrieval by means of information visualization and relevance feedback. Methods will be discussed to localize the retrieved objects in their images.

Recorded on : 2006/04/16
Filesize : 791 MB
Length : 2h23mn
Introduction to Bioinformatics

Place: SIU2006 Conference - Antalya
Speaker: Uğur Sezerman - Assistant Professor, Sabancı Üniversitesi; Attila Gürsoy - Assistant Professor, Koç Üniversitesi
Biotechnology today require interdisicplinary approaches to come up with novel solutions to existing problems. The goal of this tutorial is to bring people together from different backgrounds including Biology, Bioinformatics, Statistics and Signal Processing to open new research areas and promote new collaborations in challenging problems in genetics. The tutorial is geared towards Signal Processing community so we will start with giving definitions of the jargon used in biology (DNA , gene, protein,...)to give an idea about the domain of the problem. Next, we will cover the general concepts of bioinformatics and then focus on different applications of signal and image processing techniques to Bioinformatics. Topics will include: Basic Definitons of Biology, What is Bioinformatics, Signal processing and statistical approaches to bioinformatics problems, Digital communications approaches to gene sequence estimation, Image and signal processing for microarray data, Data mining and pattern recognition; signal extraction, Models for cellular and inter-cellular signaling

Recorded on : 2006/04/16
Filesize : 1.52 GB
Length : approx. 4:30 hrs
Sequential Inference using Particle Filters

Place: SIU2006 Conference - Antalya
Speaker: Dr. Mahesan Niranjan - Professor, University of Sheffield
Many modern signal processing problems involve systems that are nonlinear and nonstationary. Data-driven models that are based on powerful function approximation methods such as neural networks have been applied with demonstrable success to these problems. Nonstationarity imposes a particular difficulty in these settings because regularisation techniques such as cross validation can be inapplicable. This tutorial will address sequential estimation techniques that are useful in nonlinear and nonstationary environments. It will use a Bayesian dynamical systems approach and will introduce concepts and algorithms involving the extended Kalman filter (EKF) and powerful variants of it. Starting from the EKF, we will review more recent developments in sequential Markov Chain Monte Carlo (Particle Filters), and explore their application in a number of practical examples taken from speech signal processing and an image processing problem in microarray gene expression analysis.

Recorded on : 2006/04/16
Filesize : 635 MB
Length : 1h57mn
TV recordings for 'Teknoloji kanali'

There have been three demo for 'Teknoloji kanali' channel:
1- 3 dimensional scanner demo
2- Pack2000 demo
3- Object recognition demo

Recorded on : 2005/10/11
Filesize : 17.3 MB
Length : 20:37 mn
Aytul Ercil' s interview on 'Teknoloji kanali'

Aytul Ercil explaining VPA projects on channel 'Teknoloji kanali', program 'teknoloji hatti'

Recorded on : 2005/09/13
Filesize : 8.20 MB
Length : 12:13 mn
Robert Duin and Dick de Ridder seminars

There have been a seminar on June 27-29 by Robert Duin ve Dick de Ridder in Sabanci University in MDBF building arranged by VPA.
1. Summary PR, PRTools software
2. Classifiers, performance estimation

3. High-throughput data: clustering and marker gene selection
4. High-throughput data: feature selection and classification

5. Weak classifiers, bias-variance trade off
6. Boosting

Related links here and here

Recorded on : 2005/06/27
Filesize : 3.40 GB
Length : approx. 12 hr
S. Geman 's seminar

S. Geman gave a seminar about 'Four Arguments in Support of Hierarchy in Computer Vision' in Sabanci University.

Recorded on : 2005/05/05
Filesize : 752 MB
Length : 1:10 hr
CARIA in news

This is a program about CARIA project shown on 'CNN Turk' on 2005/02/02 19:45

Recorded on : 2005/02/02
Filesize : 3.62 MB
Length : 3 mn
For more information about these movies and others not cited here, you may contact us.

Home Back