Averaged stochastic optimization for medical image registration based on variance reduction

Wei Sun*, Dirk H.J. Poot, Xuan Yang, Wiro J. Niessen, Stefan Klein

*Corresponding author for this work

    Research output: Chapter in Book/Conference proceedings/Edited volumeConference contributionScientificpeer-review

    1 Citation (Scopus)

    Abstract

    In image registration the optimal transformation parameters of a given transformation model are typically obtained by minimizing a cost function. Stochastic gradient descent (SGD) is an efficient optimization algorithm for image registration. In SGD optimization, stochastic approximations of the cost function derivative are used in each iteration to update the transformation parameters. The stochastic approximation error leads to large variance in the parameters. To enforce convergence nonetheless, SGD methods are typically implemented in combination with a gradually decreasing update step size. However, selecting a proper sequence of step sizes is a major challenge in practice. An alternative strategy in numerical optimization is to use a constant step size and enforce convergence by averaging the parameters obtained by SGD over several iterations. It was proven mathematically that the highest possible rate of convergence is achieved in this way. Inspired by this work, we propose an averaged SGD (Avg-SGD) method for efficient image registration. In the Avg-SGD approach, a constant step size is used, in combination with an exponentially weighted iterate averaging scheme. Experiments on 3D lung CT scans demonstrate the effectiveness of the Avg-SGD method in terms of convergence rate, accuracy and precision.

    Original languageEnglish
    Title of host publicationBiomedical Image Registration - 8th International Workshop, WBIR 2018, Proceedings
    EditorsS. Klein, S. Sommer, S. Durrleman, M. Staring
    PublisherSpringer
    Pages69-79
    Volume10883 LNCS
    ISBN (Print)978-331992257-7
    DOIs
    Publication statusPublished - 1 Jan 2018
    Event8th International Workshop on Biomedical Image Registration, WBIR 2018 - Leiden, Netherlands
    Duration: 28 Jun 201829 Jun 2018

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volume10883 LNCS
    ISSN (Print)0302-9743
    ISSN (Electronic)1611-3349

    Conference

    Conference8th International Workshop on Biomedical Image Registration, WBIR 2018
    Country/TerritoryNetherlands
    CityLeiden
    Period28/06/1829/06/18

    Fingerprint

    Dive into the research topics of 'Averaged stochastic optimization for medical image registration based on variance reduction'. Together they form a unique fingerprint.

    Cite this