INFORMAZIONI SU

Image Processing: Computer Vision (the English translation of "Elaborazione delle immagini: visione")

Image Processing: Computer Vision (traduzione in inglese del programma di Elaborazione delle immagini: visione) - cdl magistrale in Ingegneria Elettronica

Teacher

prof. Andrea FUSIELLO

Credits

6 CFU

Language

Italian

Objectives

The course aims at providing the student with the theoretical and practical tools to tackle the problem of recovering the 3D structure of a scene starting from its 2D projections: the images. This can be seen as the inverse of Computer Graphics,  where images are rendered starting from a geometric description of the scene. The focus will be on the pinhole camera model and on the  geometric relationships among multiple projections of a scene.

Acquired skills

The student should reach the level adequate to performing the following tasks.

- Understanding of the imaging process.
- Geometrical methods for recovering three-dimensional shape from images.
- Geometrical methods for image processing and analysis.
- Geometrical methods for camera orientation.
- Image features detection and matching.

Lectures and exercises (topics and specific content)

Introduction: introduction to the course and to the discipline (2 hours).
Fundamentals of Mathematics: linear algebra, projective geometry, robust regression methods (6 hours).
Principles of imaging: geometry, photometry, optics (4 hours).
Camera model and calibration: pinhole model of the camera, calibration  (4 hours).
Stereopsis and epipolar geometry: general principle, triangulation, epipolar geometry, rectification (6 hours).
Matching: two-views matching, active methods, multi-view stereo (6 hours).
Structure and motion: structure and motion reconstruction, essential matrix, multi-view case; feature detection  (6 hours).
Optical flow:  motion field and optical flow (4 hours).
Orientation problems: relative orientation, absolute orientation, ICP, exterior orientation (4 hours).
Uncalibated reconstruction and autocalibration: projective reconstruction, euclidean upgrade, autocalibration (4 hours).
Planar scene: collineations and mosaicing: collineation induced by a plane, parallax, image mosaicing (4 hours).
Labs (10 hours).

References

- A. Fusiello, Visione Computazionale

Type of exam

Oral

Additional material or information on line