The tutorials organized at SCIA 2013 are held on Monday, June 17, at Aalto University Otaniemi campus, Computer Science and Engineering building, hall T2. See the Venue page for travel instructions. The tutorials are:

Matti PietikäinenGuoying Zhao

Matti Pietikäinen: Professor, Department of CSE, University of Oulu

Guoying Zhao: Academy research fellow, Department of CSE, University of Oulu

Time: 9:00-12:00


Texture is a fundamental property of images and videos. It can be seen almost anywhere. In recent years, very discriminative and computationally efficient texture descriptors based on local binary patterns (LBP) have been proposed, which has led to significant progress in applying texture methods to various computer vision problems. The focus of the research has also broadened from 2D textures to 3D textures and spatiotemporal (dynamic) textures. Due to this progress the application areas of texture analysis are now covering such important fields as biomedical image analysis, biometrics, face and activity analysis, human-computer interaction, industrial inspection, remote sensing, video analysis, and visual surveillance.

Lectured by pioneers of LBP, this tutorial presents effective image and video descriptors based on the highly popular LBP operator and its recent variants. Part I introduces the basic LBP operator in spatial and spatiotemporal domains. Part II provides an overview of some recent LBP variants which improve the discriminative power and robustness of the original LBP. Part III presents examples of using LBP and its variants in important image and video analysis problems and applications. Part IV deals with applications of LBP in facial image analysis. Part V concludes the tutorial and presents some directions for future research.

Tülay Adalı

Professor, Department of CSEE, UMBC.

IEEE SPS Distinguished Lecturer

Time: 13:00-16:00


Data-driven methods are based on a simple generative model and hence can minimize the assumptions on the nature of data. They have emerged as promising alternatives to the traditional model-based approaches in many applications where the underlying dynamics are hard to characterize. Independent component analysis (ICA), in particular, has been a popular data-driven approach and an active area of research. Starting from a simple linear mixing model and imposing the constraint of statistical independence on the underlying components, ICA can recover the linearly mixed components subject to only a scaling and permutation ambiguity. It has been successfully applied to numerous data analysis problems in areas as diverse as biomedicine, communications, finance, geophysics, and remote sensing.

This tutorial reviews the fundamentals and properties of ICA, and provides a unified view of two main approaches for achieving ICA, those that make use of non-Gaussianity and sample dependence. Then, the generalization of ICA for analysis of multiple datasets, independent vector analysis (IVA), is introduced and the connections between ICA and IVA are highlighted, in particular in the way both approaches make use of signal diversity. Several key problems for achieving a successful decomposition, such as matrix optimization and density matching are discussed as well both for ICA and IVA.

The second half of the tutorial presents examples of the application of ICA and IVA to medical image analysis and to fusion of multiple imaging modalities.