Using multiscale top points in image matching

Bram Platel, Frans Kanters, Luc Florack, Evguenia Balmachnova

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

16 Citations (Scopus)


Earth Mover's Distance (EMD) reflects the minimal amount of work that must be performed to transform one distribution into the other by moving "distribution mass" around. Very efficient algorithms for calculating the EMD exist. In our experiments we have used an algorithm proposed by Rubner et al. Without going into details about the EMD and its implementation an intuitive view of the EMD is as follows:Given two distributions, one can be seen as a mass of earth properly spread in space, the other as a set of holes in that same space. Assuming that there is always at least as much earth as needed to fill all the holes (this can be achieved by switching what is called earth or holes if necessary), the EMD gives a measure of the least amount of work needed to fill the holes with earth. One unit of work corresponds to transporting a unit of earth by a unit distance.This means that apart from the distributions we want to match (e.g. the top points) a weight distribution} and a distance measure are important.
Original languageEnglish
Title of host publicationProceedings IEEE Conference on Image Processing (ICIP 2004, Singapore, October 24-27, 2004)
Place of PublicationPiscataway, NJ
PublisherInstitute of Electrical and Electronics Engineers
ISBN (Print)0-7803-8554-3
Publication statusPublished - 2004
Event11th IEEE International Conference on Image Processing (ICIP 2004) - Singapore, Singapore
Duration: 24 Oct 200427 Oct 2004
Conference number: 11


Conference11th IEEE International Conference on Image Processing (ICIP 2004)
Abbreviated titleICIP 2004


Dive into the research topics of 'Using multiscale top points in image matching'. Together they form a unique fingerprint.

Cite this