A data-driven learning approach to image registration

Handling large displacement optical flow is a remarkably arduous task. For instance, standard coarse-to-fine techniques often struggle to adequately deal with moving objects whose motion exceeds their size. Here we propose a learning approach to the estimation of large displacement between two non-c...

Full description

Bibliographic Details
Main Author: Mustafa, Mohammad A.R.
Format: Thesis (University of Nottingham only)
Language:English
Published: 2016
Online Access:https://eprints.nottingham.ac.uk/33723/
_version_ 1848794689892777984
author Mustafa, Mohammad A.R.
author_facet Mustafa, Mohammad A.R.
author_sort Mustafa, Mohammad A.R.
building Nottingham Research Data Repository
collection Online Access
description Handling large displacement optical flow is a remarkably arduous task. For instance, standard coarse-to-fine techniques often struggle to adequately deal with moving objects whose motion exceeds their size. Here we propose a learning approach to the estimation of large displacement between two non-consecutive images in a sequence on the basis of a learning set of optical flows estimated a priori between different consecutive images in the same sequence. Our method refines an initial estimate of the flow field by replacing each displacement vector by a linear combination of displacement vectors at the center of similar patches taken from a code-book built from the learning set. The key idea is to use the accurate flows estimated a priori between consecutive images to help improve the potentially less accurate flows estimated online between images further apart. Experimental results suggest the ability of a purely data-driven learning approach to handle fine scale structures with large displacements.
first_indexed 2025-11-14T19:20:11Z
format Thesis (University of Nottingham only)
id nottingham-33723
institution University of Nottingham Malaysia Campus
institution_category Local University
language English
last_indexed 2025-11-14T19:20:11Z
publishDate 2016
recordtype eprints
repository_type Digital Repository
spelling nottingham-337232025-02-28T13:29:15Z https://eprints.nottingham.ac.uk/33723/ A data-driven learning approach to image registration Mustafa, Mohammad A.R. Handling large displacement optical flow is a remarkably arduous task. For instance, standard coarse-to-fine techniques often struggle to adequately deal with moving objects whose motion exceeds their size. Here we propose a learning approach to the estimation of large displacement between two non-consecutive images in a sequence on the basis of a learning set of optical flows estimated a priori between different consecutive images in the same sequence. Our method refines an initial estimate of the flow field by replacing each displacement vector by a linear combination of displacement vectors at the center of similar patches taken from a code-book built from the learning set. The key idea is to use the accurate flows estimated a priori between consecutive images to help improve the potentially less accurate flows estimated online between images further apart. Experimental results suggest the ability of a purely data-driven learning approach to handle fine scale structures with large displacements. 2016-07-19 Thesis (University of Nottingham only) NonPeerReviewed application/pdf en arr https://eprints.nottingham.ac.uk/33723/1/Thesis%20Mohammad%20A%20R%20Mustafa%20ID%204151693.pdf Mustafa, Mohammad A.R. (2016) A data-driven learning approach to image registration. PhD thesis, University of Nottingham.
spellingShingle Mustafa, Mohammad A.R.
A data-driven learning approach to image registration
title A data-driven learning approach to image registration
title_full A data-driven learning approach to image registration
title_fullStr A data-driven learning approach to image registration
title_full_unstemmed A data-driven learning approach to image registration
title_short A data-driven learning approach to image registration
title_sort data-driven learning approach to image registration
url https://eprints.nottingham.ac.uk/33723/