Camera Geometry
After a brief introduction to image acquisition and light reflection, the building blocks of modern cameras will be surveyed, along with geometric camera modelling. Several camera models, like pinhole and weak-perspective camera model, will subsequently be presented, with the most commonly used camera calibration techniques closing the lecture.
Should you require access to the resource, please contact the author directly.
Image Acquisition
This lecture overviews that has many applications in computer vision, digital imaging, social media and media production. It covers the following topics in detail: Light Reflection, Camera Structure, Camera Lens, Sensor Technologies, Image Digitization, Image Corrections, Image File Formats, Scanners, Image noise.
Should you require access to the resource, please contact the author directly.
Digital Images and Videos
This lecture overviews various digital image types: 2D images, 3D images (videos, medical volumes, hyperspectral images). Multichannel images, e.g., colour and multispectral images and colour theory come next. Spatial frequency content and image sampling are presented. Spatiotemporal frequency content and video sampling are presented.
Should you require access to the resource, please contact the author directly.
Stereo and Multiview Imaging
Stereoscopic and multiview imaging will be explored in depth, as they have tremendous applications in many applications, ranging from autonomous car/drone/robot/vessel vision to Surveying Engineering to Medical Imaging. Focus will be mainly on stereoscopic vision, geometry and camera technologies. Epipolar Geometry, Essential and Fundamental matrix, Stereo rectification are detailed. Subsequently, the main methods of disparity estimation and 3D scene reconstruction from stereoscopic video will be described, together with feature search and matching. Multiview imaging will be overviewed.
Should you require access to the resource, please contact the author directly.
An Introduction to PAC-Bayesian Analysis
This resource corresponds to 9th video from the AI Excellence Lecture Series.
PAC-Bayesian Analysis is a framework in machine learning and statistics that combines ideas from the Probably Approximately Correct (PAC) learning framework and Bayesian probability theory. It is used to analyse the generalisation performance of machine learning algorithms.