Transformational approach for alignment-free image matching applications

Komal Komal, Nandita Bhattacharjee, David Albrecht, Bala Srinivasan

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

    1 Citation (Scopus)

    Abstract

    One of the fundamental steps in computer vision and image processing applications is the alignment of images before matching. A number of alignment methods based on features points, pixel information and transformation parameters have been proposed in the literature. All these methods either suffer from accurate detection of feature points or are computationally expensive that makes the entire matching process time consuming. The fast matching of images performed without alignment can benefit a number of image matching applications. In this paper, a transformational approach is proposed that identifies whether two images are similar or not without performing an alignment. The matching without alignment reduces the computation time of the entire matching process. The effectiveness of the proposed approach has been demonstrated on UMD logo dataset. The robustness of the proposed approach to additive noise shows that high accuracy can be achieved even when the noise is as high as 20%. Furthermore, the computation time is improved by decreasing the number of projections without affecting the matching accuracy. The experiments demonstrate that the proposed method is not only efficient but also highly accurate as compared to existing approaches.

    Original languageEnglish
    Title of host publicationThe 15th International Conference on Advances in Mobile Computing and Multimedia, December 04-06, 2017, Salzburg, Austria
    Subtitle of host publicationMoMM 2017 Proceedings
    EditorsEric Pardede, Pari Delir Haghighi, Ivan Luiz Salvadori, Matthias Steinbauer, Ismail Khalil, Gabriele Anderst-Kotsis
    Place of PublicationNew York NY USA
    PublisherAssociation for Computing Machinery (ACM)
    Pages49-57
    Number of pages9
    ISBN (Print)9781450353007
    DOIs
    Publication statusPublished - 4 Dec 2017
    EventInternational Conference on Advances in Mobile Computing and Multimedia 2017 - Salzburg, Austria
    Duration: 4 Dec 20176 Dec 2017
    Conference number: 15th
    http://www.iiwas.org/conferences/momm2017/

    Conference

    ConferenceInternational Conference on Advances in Mobile Computing and Multimedia 2017
    Abbreviated titleMoMM 2017
    CountryAustria
    CitySalzburg
    Period4/12/176/12/17
    Internet address

    Keywords

    • Alignment-free
    • Computation time
    • Logos
    • Precision-Recall
    • Radon transform

    Cite this