Course content

The course is given every second year, next time in autumn 2023. The course builds on the basic knowledge gained from the robotic vision course TTK4255 with a focus on motion estimation, non-linear estimation methods and object recognition, the underlying theoretic foundation, and implementation skills. The course is given in English.

Learning outcome

Scientific Contents:

* optimization, estimation, and uncertainty models for robotic vision systems

* motion estimation, optical flow, scene flow

* scene and place recognition

* dense, semi-dense, sparse matching

* shape priors and estimation

* pose estimation and tracking

* semantic scene understanding

* statistical models and uncertainties

* segmentation and fitting using probabilistic methods

* correspondence and pose consistency

* recognition by relations between templates

SKILLS:

* Proficiency in linear and non-linear estimation methods on image and video data, statistics, and probabilistic methods

* Proficiency in critical thinking and evaluation of disadvantages of different methods and approaches to make a qualified choice of methods for a given system

* Proficiency in designing motion estimation, and object recognition systems

GENERAL COMPETENCE:

* Skills in applying this knowledge and proficiency in new areas and completing advanced tasks and projects

* Skills in communicating extensive independent work, and mastering the technical terms of advanced visual perception techniques

* Ability to contribute to innovative thinking and innovation processes

Learning methods and activities

Study groups and optional problem sets. Project with report.

Recommended previous knowledge

TTK4255 Robotic Vision

TTK4250 Sensor Fusion

Required previous knowledge

TTK4115 Linear Systems Theory or similar.

Course materials

A collection of papers, which will be given at the beginning of the semester.