Seminar

Prerequisites

This course requires knowledge of linear algebra, basic calculus, and basic probability. Previous experience with computer vision, pattern recognition, image processing and/or computer graphics will be helpful.

Course Objectives

The ubiquity of digital cameras and the internet, coupled with advances in computer vision and graphics, have been the catalysts of computational photography, an emerging field with the goal of going beyond the limitations of conventional photography by proposing new ways in which photographs can be captured, manipulated, and organized.
The general goal of this seminar is to learn how, by using computational techniques and alternative camera design, computational photography produces a richer representation of our visual world. In particular, in this course, we will study ways of manipulating and combining photographs and videos to enhance the photography experience, by touching upon areas that go from mathematical models of light to advanced photography systems and techniques.
Popular algorithms will be presented, which enable the development of image analysis and synthesis tools, needed to render novel images on the computer, from previously acquired samples of the real world (images or videos).

Textbook

Notes and reading material will be distributed. We will not rely on a textbook, although I recommend the free, online:

Computer Vision: Algorithms and Applications, by Richard Szeliski, Springer, 2011 

Other optional books are listed here:

• Computational Photography: Mastering New Techniques for Lenses, Lighting, and Sensors, by Ramesh Raskar, and Jack Tumblin, http://web.media.mit.edu/~raskar/photo/ (to be published soon)
• The Art and Science of Digital Compositing, Second Edition, by Ron Brinkmann, Morgan Kaufmann, 2008
• Fundamentals of Computer Graphics, Third Edition, by Peter Shirley, Michael Ashikhmin, and Steve Marschner, A K Peters Ltd, 2009
  Photography, 10th Edition, by Barbara London, John Upton, Jim Stone, Pearson, 2011
• Digital Image Processing, Third Edition, by Rafael C. Gonzalez, Richard E. Woods, Prentice Hall, 2008
• Computer Vision: A Modern Approach, by David A. Forsyth and Jean Ponce, Prentice Hall, 2003

Prerequisites

This course requires knowledge of linear algebra, basic calculus, and basic probability. Previous experience with computer vision, pattern recognition, image processing and/or computer graphics will be helpful.

Course Objectives

The ubiquity of digital cameras and the internet, coupled with advances in computer vision and graphics, have been the catalysts of computational photography, an emerging field with the goal of going beyond the limitations of conventional photography by proposing new ways in which photographs can be captured, manipulated, and organized.
The general goal of this seminar is to learn how, by using computational techniques and alternative camera design, computational photography produces a richer representation of our visual world. In particular, in this course, we will study ways of manipulating and combining photographs and videos to enhance the photography experience, by touching upon areas that go from mathematical models of light to advanced photography systems and techniques.
Popular algorithms will be presented, which enable the development of image analysis and synthesis tools, needed to render novel images on the computer, from previously acquired samples of the real world (images or videos).

Textbook

Notes and reading material will be distributed. We will not rely on a textbook, although I recommend the free, online:

Computer Vision: Algorithms and Applications, by Richard Szeliski, Springer, 2011 

Other optional books are listed here:

• Computational Photography: Mastering New Techniques for Lenses, Lighting, and Sensors, by Ramesh Raskar, and Jack Tumblin, http://web.media.mit.edu/~raskar/photo/ (to be published soon)
• The Art and Science of Digital Compositing, Second Edition, by Ron Brinkmann, Morgan Kaufmann, 2008
• Fundamentals of Computer Graphics, Third Edition, by Peter Shirley, Michael Ashikhmin, and Steve Marschner, A K Peters Ltd, 2009
  Photography, 10th Edition, by Barbara London, John Upton, Jim Stone, Pearson, 2011
• Digital Image Processing, Third Edition, by Rafael C. Gonzalez, Richard E. Woods, Prentice Hall, 2008
• Computer Vision: A Modern Approach, by David A. Forsyth and Jean Ponce, Prentice Hall, 2003

Thursday April 12

• Introduction
• The language of images and its relationship with the written language
• The meaning as a content
  • Models for meaning encoding and transmission
  • The context and the semiological level of the meaning
• The meaning as an interpretation
  • The context and the interpreter
• The images creation process and its influence on the content
• Computational approaches to the description of the multimedial meaning

Friday April 13

• Symbolic methods for meaning description
• Concepts
  • Linguistic concepts
  • Mathematical Theory of concepts
  • Visual concepts
• Informal symbolic methods
  • Annotations
  • MPEG-7
  • Annotation with agent-action structure
• Informal symbolic methods
  • First order logic
  • Deduction
  • Formal semantic

Monday April 16

• Logic of descriptions
  • Syntax
  • Semantic
• Uncertainty: probability and fuzzy logic
• Fuzzy description logic

• Ontologies
  • What an ontology is?
  • Open or closed world?
  • An example of multimedial ontology: BOEMIE
  • Fuzzy ontologies for multimedia data.

Wednesday April 18

• Non-symbolic methods
• Similarity hypothesis
• Preattentive and attentive similarity
• Distances in the features space
• Non-metric distances: "Tversky and Tversky fuzzy"
• Pro-semantic systems: classifiers as relationships amongs images and the surrounding text
• Emergent semantic: user interaction
• Use of context of reading
• What is nowadays the status of the semantic?

Thursday April 26

• Lecture 1 - Image search: the problem
• Lecture 2 - Features 1: spatio-temporal-color, first things first
• Lecture 3 - Features 2: measurements & invariance, desireable properties

Friday April 27

• Lecture 4 - Features 3: descriptors, to grasp the image
• Lecture 5 - Detection 1: words, lots of them
• Lecture 6 - Detection 2: to context or not to context

Wednesday May 2

• Lecture 7 - Image search: making it work: computation and samples
• Lecture 8 - Video search: performance, what can be done & what not?

Giovedì 26 Aprile

• Lecture 1 - Image search: the problem
• Lecture 2 - Features 1: spatio-temporal-color, first things first
• Lecture 3 - Features 2: measurements & invariance, desireable properties

Venerdì 27 Aprile

• Lecture 4 - Features 3: descriptors, to grasp the image
• Lecture 5 - Detection 1: words, lots of them
• Lecture 6 - Detection 2: to context or not to context

Mercoledì 2 Maggio

• Lecture 7 - Image search: making it work: computation and samples
• Lecture 8 - Video search: performance, what can be done & what not?

Giovedì 12 Aprile

• Introduzione
• Il linguaggio delle immagini e la relazione con il linguaggio scritto
• Il significato come contenuto
  • Modelli di codifica e trasmissione del significato
  • Il contesto ed il livello semiologico del significato
• Il significato come interpretazione
  • Il contesto e l'interprete
• Il processo di creazione delle immagini e la sua influenza sul contenuto
• Gli approcci computazionali alla descrizione del significato multimedia

Venerdì 13 Aprile

• Metodi simbolici di descrizione del significato
• Concetti
  • Concetti linguistici
  • Teoria matematica dei concetti
  • Concetti visuali
• Metodi simbolici informali
  • Annotazioni
  • MPEG-7
  • Annotazione con struttura agente-azione
• Metodi simbolici informali
  • Logica del primo ordine
  • Deduzione
  • Semantica formale

Lunedì 16 Aprile

• Logica delle descrizioni
  • Sintassi
  • Semantica
• Incertezza: probabilità e logica Fuzzy
• Logica delle descrizioni fuzzy

• Ontologie
  • Cos'è una ontologia?
  • Mondo aperto o mondo chiuso?
  • Esempio di ontologia multimedia: BOEMIE
  • Ontologie fuzzy per dati multimedia

Mercoledì 18 Aprile

• Metodi non simbolici
• L'ipotesi della similarità
• Similarità preattentiva e attentiva
• Distance nello spazio delle caratteristiche
• Distanze non-metriche: "Tversky e Tversky fuzzy"
• Sistemi "prosemantici": uso di classificatori come caratteristiche relazioni tra le immagini ed il testo che le circonda
• Semantica emergente: l'interazione con l'utente
• Uso del contesto di lettura
• Qual'è oggi lo stato della semantica?

1. Introduction to intrusion detection. Slides.
   • Taxonomy: net/host
   • Host IDS: early approaches
   • Network IDS: feature extraction
   • Signature-based IDS:
2. Anomaly detection for network security.  Slides.
   
   • Early approaches (packet headers, PAYL, Anagram)
   • Payload based approach
   • applications (ReMIND experiments, SCADA systems)
3. Classification for network security. Slides.
   • Early approaches
   • Why it is difficult to use classification for IDS
   • behavioral classification
   • automatic signature generation
4. Attacks against learning algorithms. Slides.
   • attack taxomony
   • attacks against anagram
   • attacks against automatic signature generation

1. Introduction to intrusion detection. Slides.
   • Taxonomy: net/host
   • Host IDS: early approaches
   • Network IDS: feature extraction
   • Signature-based IDS:
2. Anomaly detection for network security. Slides.
   • Early approaches (packet headers, PAYL, Anagram)
   • Payload based approach
   • applications (ReMIND experiments, SCADA systems)
3. Classification for network security. Slides.
   • Early approaches
   • Why it is difficult to use classification for IDS
   • behavioral classification
   • automatic signature generation
4. Attacks against learning algorithms. Slides.
   • attack taxomony
   • attacks against anagram
   • attacks against automatic signature generation