Talks of Luc Brun and Lorenzo Bruzzone – 4th of July

The defense of Yanwei Cui is going to be held on tuesday, 4th of July. We are glad to welcome  Luc Brun and Lorenzo Bruzzone  (members of the PhD commitee) for 2 talks from 9:30 to 11:30 for two talks (details below)
You’re more than welcome to attend.
Luc Brun
Normandy University, ENSICAEN, CNRS, Caen, France

Title: Graph Edit Distance: Basics and trends

Defining a metric between objects is a basic step of any pattern recognition algorithm. Using graphs, this notion of distance is not straightforward. Among the different distances between graphs that one may imagine, the Graph Edit Distance has progressively become a standard tool within the structural pattern recognition framework. Indeed, this distance allows to take into account fine differences between graphs, may be easily tuned and may satisfy all the axioms of a distance. Basically, the most common definition of the graph edit distance is based on the notion of edit path. An edit path between two graphs
G1 and G2 is a sequence of node/edge removal/substitution or insertion
operations transforming G1 into G2. Each edit path may be associated to a cost hence defining the Graph Edit Distance between G1 and G2 as the minimal cost of all edit paths between these two graphs.
In this talk, after the introduction of some basic definitions and concepts we will review the main families of methods used to compute the graph edit distance. This talk ends on recent and efficient methods still under development.
Lorenzo Bruzzone
Remote Sensing Laboratory, Dept. of Information Engineering and Computer Science, University of Trento, Italy
Title: Current scenario and challenges in the analysis of multitemporal remote sensing images

In the last decade a large number of new satellite remote sensing missions have been launched resulting in a dramatic improvement in the capabilities of acquiring images of the Earth surface. This involves an enhanced possibility to acquire multitemporal images of large areas of the Earth surface, with improved temporal and spatial resolution with respect to traditional satellite data. Such new scenario significantly increases the interest of the remote sensing community in the multitemporal domain, requiring the development of novel data processing techniques and making it possible to address new important and challenging applications. The potentials of the technological development are strengthen from the increased awareness of the importance of monitoring the Earth surface at local, regional and global scale. Assessing, monitoring and predicting the dynamics of land covers and of antrophogenic processes is at the basis of both the understanding of the problems related to climate changes and the definition of politics for a sustainable development. Nonetheless, the properties of the images acquired by the last generation sensors pose new methodological problems that require the development of a new generation of methods for the analysis of multitemporal images and temporal series of data.
After a general overview of the problems related to the analysis of multitemporal images and time series of data, this talk will focus on the very important problem of automatic change detection between multitemporal images. The development and the use of effective automatic techniques for change detection are of major importance in many of the above-mentioned application scenarios. The increased geometrical resolution of multispectral and SAR sensors, the increased revisit time of high resolution systems, and the expected availability of time series of hyperspectral images in the near future result in many different methodological problems as well as in very important new possible applications. The talk will address these problems by pointing out the state of the art and the most promising methodologies for change detection on images acquired by the last generation of satellite sensors. Examples of the use of change-detection approaches in operative scenarios will be provided.

Leave a Reply

Your email address will not be published.