Speakers

Download lectures


Speaker: Prof. Sebastiano BATTIATO
Lecture TITLE: Image Forensics: Recent Trends and Challenges
Lecture Abstract: The widespread adoption of digital content over traditional physical media has given rise to a number of new information security challenges. Digital content can be altered, falsified, and redistributed with relative ease by adversaries.
This has important consequences for governmental, commercial, and social institutions that rely on digital information. The pipeline which leads to ascertain whether an image has undergone to some kind of forgery leads through the following steps: determine whether the image is "original" and understand/reconstruct its past history.
Although the field of information forensics is still young, many forensic techniques have been developed to detect forgeries, identify the origin, and trace the processing history of digital multimedia content. This course provides an overview of information forensics research and related applications.
Also we examine the device-specific fingerprints left by digital image and video cameras along with forensic techniques used to identify the source of digital multimedia files. Finally, an overview of the recent trends and evolution will be provided.

Contact: IPLAB@CT: www.dmi.unict.it/~iplab, Home: www.dmi.unict.it/~battiato, email battiato@dmi.unict.it

Speaker: Prof. Virginio CANTONI
Lecture TITLE: Eye-tracking systems, research and applications
Lecture AbstractWYSIWYG: What you see is what you get. An eye tracker is a device for measuring either eye positions and eye movement or the point of gaze (where one is looking).

Following the “strong eye-mind hypothesis” (dated in the ’80) when a subject looks at a word or object, he or she also thinks about.  Current consensus is that visual attention is always slightly ahead of the eye: as attention moves to a new target, the eyes saccade mechanism is prompted to follow.

In this connection there is an increasing interest in eye tracking and its impact in academic and scientific research today is very strong. The eye tracker is non-intrusive and the information gathered can be exploited in a variety of researches that analyzes human behavior and patterns, such as psychology, neuroscience, psycholinguistics, training and learning.

In HCI the gaze-contingency paradigm is a general framework for techniques allowing a computer screen display to change in function depending on where the viewer is looking: the eye-tracker is used as an input device and the computer interfaced responds to the observer’s fixations and interacts with him.

The gaze-contingent technique is then the basis of a variety of different applications, such as: web usability (eye tracking reveals consumers ‘attention’ as they are exposed to an advertisement or test information);market analysis (print advertisements and mail-outs, TV spots and videos, point-of-sale, online marketing, etc.); education and e-learning (in prevalence for distance education being able to assess where students are looking and also their efforts in order to improve web usability for online sessions); game design (knowing where a player looks or doesn’t look game developers can see their games literally from the eyes of their users); package design (examining the visual behavior of a consumer while interacting with a target package, evaluating distinctiveness, attractiveness and the tendency of the package); automotive engineering (recording and analyzing driver attention for road-safety and driver distraction studies, assessing the effectiveness of signage, for dashboard and instrumentation design).

Finally eye tracking is used in communication systems for the benefit of users with disabilities. Let us quote a network of excellence on the subject: COGAIN (COmmunication by GAze Interaction). Started in 2004, the COGAIN project, supported by the European Commission's, now became an association for the development of applications that allow the user to speak, send e-mail, browse the Internet, up-to the EyeArt program, and perform other such activities, using only their eyes.

Contact: Website http://vision.unipv.it/people/cantoni/index.html, email virginio.cantoni@unipv.it

Speaker: Prof. Rita CUCCHIARA
Lecture TITLE: Tracking from static, moving and multiple cameras: evaluations and trends
Lecture Abstract: Tracking is one of the most challenging computer vision problems, concerning the task of generating an inference about the motion of an object given a sequence of images. The ultimate goal is establishing the location of the target over time starting from the knowledge of its visual aspect and  position given in an initial frame. More than fifteen years of research resulted in an incredible plethora of different approaches, concerning different conditions on  the environment ( constrained- unconstrained, cluttered, indoor-outdoor..), the acquisition system ( static, moving, multiple cameras), the target ( single, multiple, with variable aspect…); however a satisfactory solution is still far away, in many circumstances. In the talk, we will address a nontrivial, inclusive presentation of tracking models and solutions, discussing the different parts of tracking, starting form detection, representation, prediction, data association. Then we will discuss methodologies for performance evaluations with a specific reference on single-target tracking and multiple-target tracking. Tens of different methods will be compared and evaluated Then in the second part of the talk, some state-of-the-art approaches will be presented with specific reference to both short term and long term tracking with their possible application in surveillance;  the talk will point out also some new trends in semi-supervised learning approaches for tracking, and in combined tracking & detection approaches.
Contact: Website http://imagelab.ing.unimore.it ,  email Rita.cucchiara@unimore.it

Speaker: Prof. Leila DE FLORIANI
Lecture TITLE: Computational topology tools for data analysis.
Lecture Abstract: Spatially-related digital data are being produced at a constantly increasing pace and their availability is changing the approach to science and its applications. The complexity of the data derives not only from the size of currently available data sets, which often consist of huge collections of unorganized samples, but also comes from the need to filter out relevant information from huge quantities of unimportant details. This leads to the need for computational tools that can efficiently process large sets of data and generate synthetic descriptors, which should adapt to different applications, such as classification, recognition, visualizatin, reconstruction, etc. Topology deals with qualitative geometric information, and this provides either an alternative, or at least a complementary, way to describe objects. Cognitive studies have shown that the human visual system makes an extensive use of topological information. On the other hand, the use of topology-based methods in a digital context is definitely not straightforward: most existing techniques only use a narrow set of properties (e.g., connectedness, Euler characteristic), while the computation of more powerful homological invariants, or structural descriptors, is an active research field. In many cases data analysis involves dealing with large data sets and high-dimensional domains, which lead to a number of open problems of both theoretical and computational nature which must be solved before these methods can become effective in real-world applications. The main issues in extensively using topological tools, like homology or Morse decompositions, in the applications is their high computational and storage costs, and lack of scalability with the increase of dimension.
In this talk, we will give an overview of existing approaches to topological data analysis. Then, we will focus on data structures and algorithms for the efficient computation of homology and Morse complexes on large and high-dimensional data sets by considering tools from discrete Morse theory and new hierarchical descriptors. We will discuss applications to visualization and to spatial data analysis.
Contact: Website  http://www.disi.unige.it/person/DeflorianiL/, email deflo@disi.unige.it>

Speaker: Prof. Alberto DEL BIMBO
Lecture TITLE: Human Action Recognition 

Lecture Abstract: Human actions are short task oriented body movements such as “waving a hand”, or “drinking from a bottle”. Some actions are atomic but often actions of interest have a cyclic nature such as “walking” or “running”…….  Activities instead, involve multiple people or happen in longer timeframes. Activities are often the result of a combination of actions like “taking money out from ATM” or “waiting for a bus”… Finally, we often refer to an Event as a combination of activities, usually involving more people and happening in a given context such as “a soccer match”, a “car accident” or a “fire in a wood”…….. A generic action recognition framework needs a robust representation in order to have classifiers concentrate on the real discriminant spatio-temporal features and not to get distracted by clutter or other irrelevant intra-class variations. Camera motion need to be either removed via motion compensation or with robust representations that are able to remove all the noisy features (clothing, gender, illumination, scale etc.) and preserve variability with respect to the body motion involved in different actions. With reference to these themes, this short course will address the state of the art of:

Detection and tracking in image sequences (recall)

Spatio-temporal descriptors

Action recognition

Event and anomaly detection

Contact: Website  www.micc.unifi.it/delbimbo/, email alberto.delbimbo@unifi.it


Speaker: Ing. Pierpaolo MURRIERI
Lecture TITLE: Industrial implementations of Visual Technologies for Homeland Security
Lecture Abstract: Today the Homeland Security has become a priority in many countries. The influence of a complex geopolitical scenario has influenced the movement of huge migration flows which impacts on the social equilibrium. This factor catalyzes the lasting instability due to the economic crisis. To face new menaces are necessary very effective technologies. The proliferation of cameras, both for CCTV and for special uses, gives the way to insert a greater degree of “smartness” on the field. The use of special visual technology improves the effectiveness of this cameras network. This short course gives a picture of some industrial application of visual technologies in homeland security. After a general introduction on this theme, biometrics and other application fields  are introduced as examples of use.

Contact: Website  www.selex-es.com , email pierpaolo.murrieri@selex-es.com


Speaker: Prof. Marcello PELILLO
Lecture TITLE: Creativity and Scientific Research
Lecture Abstract: In this lecture I’ll be talking about the role of creativity in scientific research. After briefly discussing the issue of whether creative processes can be implemented by machines, I’ll take the audience through a journey into the thought of the most creative thinkers of the twentieth century, including H. Poincaré, J. Hadamard, G. Polya, etc., who have devoted much of their (late) work to understanding the psychology of invention in mathematics and science. I’ll then conclude the lecture by discussing the connections between creativity and the notion of scientific progress as described by T. Kuhn and his followers.

Contact: Website http://www.dsi.unive.it/~pelillo , email: marcello.pelillo@gmail.com


Speaker: Prof. Alfredo PETROSINO
Lecture TITLE:Learning in computer vision: deep, online and related issues.
Lecture Abstract:Learning with its tolerance for imprecision, uncertainty, approximate reasoning, and partial truth in order to achieve tractability, robustness, low-cost solutions and close resemblance with human like decision-making is highly required in computer vision tasks like identification, recognition and tracking of patterns in unstructured environments. Specifically, learning multiple levels of representation and abstraction has great success in computer vision, motivated by the need to learn features, rather than hand-craft them.
The goal is to provide a high-level, broad, and rigorous overview of the algorithmic and practical aspects of learning. By the end of tutorial the attendees should have acquired enough knowledge to be able to pin-point a learning algorithm that best matches an application they face. The basic definitions of learning, and specifically deep and online, will be firstly introduced. While the goal of online learning is to make a sequence of accurate predictions given knowledge of the correct answer to previous prediction tasks and possibly additional available information, the goal of deep learning is to learn features in a 'stacked' way, realized into hierarchies that can extract multiple layers of representation. Throughout, links will be drawn between these methods and existing approaches to recognition, particularly those involving hierarchical representations. The final part of the lecture will examine the current performances obtained by feature learning approaches on a range of standard vision benchmarks, highlighting their strengths and weaknesses.

Contact: Website http://cvprlab.uniparthenope.it , email: alfredo.petrosino@uniparthenope.it


Speaker: Prof. Fabio ROLI
Lecture TITLE:ADVERSARIAL PATTERN RECOGNITION
Lecture Abstract:Pattern classifiers are currently used in several applications, like biometric recognition, spam filtering, and intrusion detection in computer networks, which are different from traditional pattern recognition tasks. The difference lies in the fact that in these applications an intelligent, adaptive adversary can actively manipulate patterns with the aim of making a classifier ineffective, namely, with the aim of evading it. Traditional pattern recognition techniques do not take into account the adversarial nature of classification problems like the ones mentioned above. One of the consequences is that the performance of standard pattern classifiers can significantly degrade when they are used in adversarial tasks.This kind of problem has been named adversarial classification, and is the subject of an emerging research field in the machine learning and pattern recognition communities. The purposes of this tutorial are: (a) to introduce the fundamentals of adversarial classification from the perspective of a designer of a pattern recognition system; (b)to illustrate the design cycle of a pattern recognition system for adversarial tasks, (c) to present the new techniques that have been recently proposed to assess performance of pattern classifiers under attack, evaluate classifiers’ vulnerabilities, and implement defence strategies that make classifiers more robust against attacks; (d) to show some applications of adversarial classification techniques to pattern recognition tasks like biometric recognition and spam filtering.

Contact: Website http://pralab.diee.unica.it/FabioRoli , email roli@diee.unica.it


Speaker: Prof. Hanan SAMET (ACM Distinguished Speaker)
Lecture TITLE:PLACE-BASED INFORMATION SYSTEMS: TEXTUAL LOCATION IDENTIFICATION AND VISUALIZATION
Lecture Abstract: The popularity of web-based mapping services such as Google Earth/Maps and Microsoft Virtual Earth (Bing), has led to an increasing awareness of the importance of location data and its incorporation into both web-based search applications and the databases that support them,In the past, attention to location data had been primarily limited to geographic information systems (GIS), where locations correspond to spatial objects and are usually specified geometrically.

However, in the web-based applications, the location data often corresponds to place names and is usually specified textually. An advantage of such a specification is that the same specification can be used regardless of whether the place name is to be interpreted as a point or a region.
Thus the place name acts as a polymorphic data type in the parlance of programming languages. However, its drawback is that it is ambiguous. In particular, a given specification may have several interpretations, not all of which are names of places. For example,``Jordan'' may refer to both a person as well as a place. Moreover, there is additional ambiguity when the specification has a place name interpretation. For example, ``Jordan'' can refer to a river or a country while there are a number of cities named ``London''.

In this talk we examine the extension of GIS concepts to textually specified location data and review search engines that we have developed to retrieve documents where the similarity criterion is not based solely on exact match of elements of the query string but instead also based on spatial proximity. Thus we want to take advantage of spatial synonyms so that, for example, a query seeking a rock concert in Tel Aviv would be satisfied by a result finding a rock concert in Herzliyah of Petach Tikva.

This idea has been applied by us to develop the STEWARD (Spatio-Textual Extraction on the Web Aiding Retrieval of Documents) system for finding documents on website of the Department of Housing and Urban Development. This system relies on the presence of a document tagger that automatically identifies spatial references in text, pdf, word, and other unstructured documents.

The thesaurus for the document tagger is a collection of publicly available data sets forming a gazetteer containing the names of places in the world. Search results are ranked according to the extent to which they satisfy the query, which is determined in part by the prevalent spatial entities that are present in the document. The same ideas have also been adapted by us to collections of news articles as well as Twitter tweets resulting in the NewsStand and TwitterStand systems, respectively, which will be demonstrated along with the STEWARD system in conjunction with a discussion of some of the underlying issues that arose and the techniques used in their implementation. Future work involves applying these ideas to spreadsheet data.
Contact: Website www.cs.umd.edu/~hjs , email: hjs@cs.umd.edu

Speaker: Ing. Mario SAVASTANO
Lecture TITLE:Jurisdictional, Legal and Ethical aspects of the security technologies
Lecture Abstract: It is a widely shared opinion that the impetuous evolution of the technology should not disregard a careful consideration of the correlated jurisdictional, legal and ethical aspects.

With particular reference to the field of "security", also because of some media pressures, the public opinion is often disoriented by the new technologies and by their potential “collateral effects”.

Apart from possible external influences, the reality is that, however, in a considerable number of “technological cases”, the boundary between technology and other issues is very thin.

In  other terms, several times, especially for the security area, the technology, the juridical issues and ethics have to cohabit in a "gray area" whose contours vary not only in function of the geopolitical area of competence but also of the evolution of the social aspects or political events that orient the public opinion.

On the basis of such premise, the object of the course proposed in the VISMAC 2014 context, will focus on some jurisdictional, legal and ethical aspects of the security technologies with particular reference to biometrics and video-analysis.

Contact: email: mario.savastano@unina.it


.


Highlights



Contact Info


University of Salerno
Via Giovanni Paolo II, nr. 132
84084 Fisciano SA, Italy

T:  (+39) 089.963334
T:  (+39) 089.963345



W:http://www.unisa.it    
E:  biplab@unisa.it



Locations of visitors to this page