Computer and robotic vision

Course objectives

In this course students will acquire fundamental knowledge in the field of computer and robotic vision with the aim to understand all the elements of today's standard engineering approach. Besides the engineering approach, we will also present Marr's complementary approach which is closer to cognitive science with the corresponding representational framework in four levels, image level, primary sketch, 2,5D representation and 3D representation.

We will also consider biological inspirations for selected standard engineering solutions which are used in practice.

Teamwork and seminar will be additional components enabling further exploration of selected topics in the field of computer and robotic vision.

Enrolment requirements and/or entry competences required for the course

-

Learning outcomes at the level of the programme to which the course contributes

  • Apply theoretical knowledge of the fundamentals of the six core disciplines and their relationship within cognitive science.
  • Critically evaluate cognitive science findings and synthesize information to be employed in a collaborative professional environment.
  • Participate in data-driven innovation projects and apply appropriate data science tools. Apply AI tools in concrete tasks and practical contexts.

Course content (syllabus)

  1. Introduction. Overview of computer vision tasks: feature extraction, segmentation, object detection, object recognition, finding object's position and orientation, visual feedback loop.
  2. Connection between biological and computer vision. Marr's framework: image, primary sketch, 2,5D sketch, 3D model.
  3. Light. Intensity, contrast, color, associated physical units and their measurement.
  4. Image representation for computer analysis. Digital image.
  5. Monocular vision. Image formation model. Perspective projection. Pinhole camera. Intrinsic and extrinsic camera parameters.
  6. Stereopsis and stereo vision. Epipolar geometry. Fundamental and essential matrix. Distance measurements.
  7. Illumination. Shadows. 3D from shading. Photometric stereo.
  8. Ridge, edge, and corner detection. Laplaceian of Gaussian and Hessian operator.
  9. The concept of scale. Multi-scale analysis.
  10. Feature extraction. Features invariant to translation and rotation.
  11. Image areas. Texture and texture analysis.
  12. mage segmentation. Artificial neural networks for image segmentation.
  13. Movement. Optical flow. Analysis of dynamic scenes.
  14. Applications: Visual quality control.
  15. Applications: Visual feedback loop and visual guidance of a robot.

Student responsibilities

he student is required to attend lectures and seminar, to prepare and to defend a seminar paper, and to pass bith written and oral exams.

Required literature

  • Richard Szeliski, "Computer Vision: Algorithms and Applications, 2nd ed."
  • David Marr, "Vision - A Computational Investigation into the Human Representation and Processing of Visual Information", The MIT Press, 2010.

Optional literature

  • Stephen E. Palmer, "Vision Science: Photons to Phenomenology", The MIT Press, 1999.