COMPUTATIONAL VISION

COMPUTATIONAL VISION

_
iten
Code
90539
ACADEMIC YEAR
2018/2019
CREDITS
6 credits during the 1st year of 10852 COMPUTER SCIENCE (LM-18) GENOVA

6 credits during the 1st year of 8733 Computer Engineering (LM-32) GENOVA

SCIENTIFIC DISCIPLINARY SECTOR
INF/01
TEACHING LOCATION
GENOVA (COMPUTER SCIENCE )
semester
1° Semester
Teaching materials

OVERVIEW

The course offers an introduction to state-of-the-art methods for visual data analysis. In particular it deals with image and video understanding.

AIMS AND CONTENT

LEARNING OUTCOMES

Students will be provided with an an overview of state-of-the-art methods for modeling and understanding the semantics of a scene. Students will get acquainted with the problem of representing the image content adaptively by means of shallow or deep computational models, then it focuses in particular on biologically-inspired hierarchical models for representing visual cues, such as discontinuity, disparity and motion. Students will also be exposed to image classification and categorization problems. Students will be involved in project activities.

AIMS AND LEARNING OUTCOMES

Students will be provided with an an overview of state-of-the-art methods for modeling and understanding the semantics of a scene. Students will get acquainted with the problem of representing the image content adaptively by means of shallow or deep computational models. Then it will address image classification and categorization problems. Possible extensions to depth and motion information will also be discussed.

Students will be involved in project activities.

PREREQUISITES

Calculus and linear algebra.

Digital image processing and machine learning principles.

Teaching methods

Theoretical classes complemented by practical activities

SYLLABUS/CONTENT

Course content

  • Introductory classes
    • Reviewing background knowledge from image processing (filters, features, histograms, color, ...) and machine learning (clustering and classification algorithms)
    • Problems formulation: image matching, image retrieval, image classification 
  • Image representations
    • Early approaches: keypoints and bag-of-keypoints
    • Sparse coding over fixed over-complete dictionaries
    • Learning adaptive dictionaries (dictionary learning)
    • Coding-pooling approaches 
    • Deep architectures
  • Additional topics: using context, dealing with temporal or depth information, data visualization issues
  • Projects and study cases

RECOMMENDED READING/BIBLIOGRAPHY

material provided by the instructors (slides and papers)

additional reference online book http://szeliski.org/Book/

TEACHERS AND EXAM BOARD

Ricevimento: Appointment by email: francesca.odone@unige.it (always specify name and surname, course name, degree name)

Exam Board

LORENZO ROSASCO (President)

FRANCESCA ODONE (President)

ALESSANDRO VERRI

NICOLETTA NOCETI

ANNALISA BARLA

LESSONS

Teaching methods

Theoretical classes complemented by practical activities

EXAMS

Exam description

  • 50% theory (oral exam)
  • 50% application (individual project+seminar)

Assessment methods

  • timely delivery of assignments
  • active participation in class and on the online students forum (aulaweb)
  • final project on a use-case (datathon-like) and presentation of the obtained results in a seminar
  • oral exam

Exam schedule

Date Time Location Type Notes
14/02/2020 09:00 GENOVA Esame su appuntamento