On stochastic gradient descent and quadratic mutual information for image registration

Abhishek Singh, Narendra Ahuja

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Mutual information (MI) is quite popular as a cost function for intensity based registration of images due to its ability to handle highly non-linear relationships between intensities of the two images. More recently, quadratic mutual information (QMI) has been proposed as an alternative measure that computes Euclidean distance instead of KL divergence between the joint and the product of the marginal densities of pixel intensities. In this paper, we examine the conditions under which QMI is advantageous over the classical MI measure, for the image registration problem. We show that QMI is a better cost function to use for optimization methods such as stochastic gradient descent. We show that the QMI cost function remains much smoother than the classical MI measure on stochastic subsampling of the image data. As a consequence, QMI has a higher probability of convergence, even for larger degrees of initial misalignment of the images.

Original languageEnglish (US)
Title of host publication2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings
Pages1326-1330
Number of pages5
DOIs
StatePublished - Dec 1 2013
Event2013 20th IEEE International Conference on Image Processing, ICIP 2013 - Melbourne, VIC, Australia
Duration: Sep 15 2013Sep 18 2013

Publication series

Name2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings

Other

Other2013 20th IEEE International Conference on Image Processing, ICIP 2013
CountryAustralia
CityMelbourne, VIC
Period9/15/139/18/13

    Fingerprint

Keywords

  • Mutual information
  • image registration

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Cite this

Singh, A., & Ahuja, N. (2013). On stochastic gradient descent and quadratic mutual information for image registration. In 2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings (pp. 1326-1330). [6738273] (2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings). https://doi.org/10.1109/ICIP.2013.6738273