Kader Pustu-Iren (M.Sc.)

Research Assistant

Email: kader.pustutibeu

Phone: +49 511 762-14704

Postal address: Welfengarten 1 B, 30167 Hannover

Address for visitors: Lange Laube 28, 30159 Hannover,

Room: 2.10

As part of a project for indexing large video collections, I am working on person recognition and visualization techniques for effective exploration. My research focus is on visual question answering for videos (VideoQA). This involves the extraction of visual information, including person/concept recognition and geolocation estimation, as well as the modeling of knowledge graphs to automatically answer questions based on the visual content of a video. My work is based on deep learning methods.

  • K. Pustu-Iren, M. Mühling, N. Korfhage, J. Bars, S. Bernhöft,  A. Hörth, B. Freisleben, R. Ewerth:
    Investigating Correlations of Inter-coder Agreement and Machine Annotation Performance for Historical Video Data
    In: International Conference on Theory and Practice of Digital Libraries (TPDL), Oslo, Norway, 2019, 107-114.

    https://arxiv.org/abs/1907.10450

  • E. Müller-Budack, K. Pustu-Iren, R. Ewerth:
    Geolocation Estimation of Photos using a Hierarchical Model and Scene Classification
    In: Proceedings of European Conference on Computer Vision (ECCV), München, 2018, 575-592.

    http://openaccess.thecvf.com/content_ECCV_2018/html/Eric_Muller-Budack_Geolocation_Estimation_of_ECCV_2018_paper.html

  • E. Müller-Budack, K. Pustu-Iren, S. Diering, R. Ewerth:
    Finding Person Relations in Image Data of News Collections in the Internet Archive
    In Proceedings of International Conference on Theory and Practice of Digital Libraries (TPDL), Porto, Portugal, 2018, 229-240. “Honorable Mention Award”

    https://arxiv.org/abs/1806.08246