A Crash Course in Geometric Computer Vision, by Prof. Andrea Fusiello

Registrations are open for "A Crash Course in Geometric Computer Vision", a 3-day course to be held from the 12th to the 14th of April at DIBRIS. The course will be held by Prof. Andrea Fusiello, invited by our MaLGA faculty members Francesca Odone and Nicoletta Noceti.

If you want to attend, please register by April 9th at the link below.

Abstract

This short but intensive course will cover the fundamentals of geometric computer vision, and provide a detailed explanation of the underlying algorithms, accompanied by clear and concise MATLAB code. The course is accessible to anyone with a linear algebra background. (No neural network will be harmed!)

The first day will cover the pinhole camera model, intrinsic and extrinsic parameters, and calibration. The calibration section will delve into the Direct Linear Transform and Sturm-Maybank-Zhang algorithms. The second day will cover absolute and exterior orientation, as well as orthogonal Procrustes analysis. Additionally, the second day will address two-view geometry, including the Fundamental matrix, homography, and essential matrix. The final day of the course will cover reconstruction from images; the second part of the day will be dedicated to adjustment of independent models, Intersection-resection, Synchronization, and Bundle adjustment.

Outline

Day 1, part 1 (09:30 - 12:30) - Room 217

The pinhole camera model, intrinsic and extrinsic parameters

Day 1, part 2 (14:30 - 17:30) - Room 217

Calibration: Direct Linear Transform a and Sturm-Maybank-Zhang algorithms

Day 2, part 1 (09:30 - 12:30) - Room 217

Absolute and exterior orientation, orthogonal Procrustes analysis

Day 2, part 2 (14:30 - 17:30) - Room 217

Two-view geometry: Fundamental matrix, homography, essential matrix

Day 3, part 1 (09:30 - 12:30) - Room 710

Reconstruction from two images, reconstruction from many images.

Day 3, part 2 (14:30 - 17:30) - Room 710

Adjustment of independent models, Intersection-resection, Synchronization, Bundle adjustment.

Bio

Andrea Fusiello received his Master’s degree in Computer Science from the University of Udine in 1994. In 1993-94 he worked within the Computer Vision Group at IRST (now FBK), in Trento. He received his PhD in Computer Engineering from the University of Trieste in 1999. As a PhD student he worked at the University of Udine. He was Visiting Research Fellow at Heriot-Watt University, Edinburgh in 1999. From 2000 to 2011 he was as Researcher and then Associate Professor at the Department of Computer Science of the University of Verona, where he taught Computer Graphics and Computer Vision. In 2012 he moved to the University of Udine, where he teaches Fundamentals of Computer Science (undergraduate) and Computer Vision (graduate), and was granted the national scientific qualification as full professor in 2013. His research covers various topics in Computer Vision, Photogrammetry and Image Analysis, with a focus on 3-D modelling/reconstruction.

If you want to attend, please register by April 9th at the link below.

The course will not be streamed online.