The participants (graduate students, Ph.D. students, young researchers) will learn about the latest research approaches in 3d color image processing, contributing to their personal and scientific development. They will be encouraged to join the IEEE and IEEE SPS. In the first week we cover advanced topics in color image acquisition and low-level vision. Lectures in the second week are concerned with application to 3D color image registration and processing. The morning lectures comprise 4 hours of teaching organized in two 90 minutes sessions with 15 minutes coee break in between. In the afternoon 3 hours of exercise accompany the morning lectures. On the last day participants can earn a grade by attending the final examination. Credit transfer is negotiated with the origin universities of the incoming students and transfer will be guaranteed, where required by the participants. In total students participate in 55 hours of teaching which corresponds to 4 hours per week in a 13 week term as it is typical in Europe. We grant 6 ECTS for successful participation
which are required in the curricula of the participating master students.
Color I: Color Appearance Models
Lecturer: Arto Kaarna (Lappeenranta University of Technology) and Heikki Kälviäinen
(Lappeenranta University of Technology), Senior member IEEE
Description: This lecture covers use of Color Appearance Models (CAM) to describe perception of colors depending on context. CAM enables modeling of colors in real environments, which extends the simple, global illumination compensation applied for example in digital cameras.
Filtering: Group Theoretical Methods and their Applications to Colour
Lecturer: Reiner Lenz (Linköping University)
Description: In the simplest form the Fourier transform decomposes a function on the circle into a linear combination of sin/cos functions. Two natural questions are the following: (A) What is so special about the trigonometric functions and (B) what is the corresponding Fourier transform on other domains like the line, the sphere or the cone? The answer to both questions will lead to new methods for colour image processing.
Color II: Content-based Image Retrieval
Lecturer: Constantin Vertan (Politehnica University), Senior Member IEEE, Secretary of
the Romanian SPS Chapter
Description: The paradigm of searching for images based on their content visual similarity has five essential components: model selection, assessing the visual content, evaluating the similarity between the image example and other images of image database, ordering images in descending order of similarity to the example and their presentation to the user. All these components are presented in this lecture.
RGB-D: Advances in RGB-D
Lecturer: Michael Ruhnke (University of Freiburg)
Description: In the last 5 years we have seen impressive improvements in the capabilities of depth cameras which provide per pixel RGB and Depth information (as in Microsoft Kinect). The combination of depth and visual information opens up new challenges and opportunities for some of the most fundamental research areas in 3D reconstruction, including 3D mapping and localization, object recognition, object tracking, and activity recognition.
Point Clouds: Point Cloud Color Segmentation with Graphical Models
Lecturer: Dietrich Paulus (University of Koblenz-Landau), Member IEEE
Description: This lecture will cover an introduction to probabilistic graphical models and their application to segmentation based on depth and color information.
Segmentation: Color Image Segmentation and Texture Characterization
Lecturer: Noël Richard (University of Poitiers), Member IEEE
Description: The most representative color image segmentation approaches will be presented, including watershed, JSEG, pyramidal and texture-based. Various texture characterization techniques will include statistical invariants, co-occurrence matrices and correlograms, etc.
Color III: Color Fractal Features and Active Contours for Color Images
Lecturer: Mihai Ivanovici (Transilvania University), Member IEEE
Description: The fundamental concepts of fractal geometry will be presented, including fractal dimension and lacunarity and approaches for estimating them and how they can be used for color texture characterization, for texture classification or image segmentation purposes. The extension to color domain of existing difusion models will be presented, and how they can be used for active contour-based image segmentation.
In addition to the lectures listed there will be activities which include a one-day workshop in collaboration with the Romanian Chapter of IEEE SPS, where the local companies acting in the field of image processing will exhibit their products and present demos of their technologies. Scientific presentations will be organized and technical discussions will be initiated. One-day mountain trip will be also organized (in the “Postavaru” mountain http://en.wikipedia.org/wiki/Post%C4%83varul_Massif).
The guest lecture on Wednesday (2h) will be given by Pierre Soille (Member IEEE; Joint Research Centre, Ispra, Italy) with the topic “Colour image morphology and application to satellite images”.