VISION/Optical flow estimation year 2025-2026
A course for IMA
and DIGIT programs
The course contains two lectures on optical flow estimation, one
tutorial work, and two practical works. Various methods are aborded:
- early methods with spectral analysis, Hough detectors, image difference
based approaches, block-matching;
- dense estimation with local and
global approaches;
- regularization, robust estimators, multi-resolution;
- supervised and unsupervised machine learning methods;
- data assimilation;
- qualitative and quantitative evaluation;
- and more!
- See also: Antoine Manzanera's VISION page
Schedule
Evaluation: submit your practical work
on Moodle 2025.
Dominique Béréziat