• English
  • Deutsch
  • Log In
    Password Login
    or
  • Research Outputs
  • Projects
  • Researchers
  • Institutes
  • Statistics
Repository logo
Fraunhofer-Gesellschaft
  1. Home
  2. Fraunhofer-Gesellschaft
  3. Konferenzschrift
  4. Learning Embeddings for Image Clustering: An Empirical Study of Triplet Loss Approaches
 
  • Details
  • Full
Options
2021
Conference Paper
Titel

Learning Embeddings for Image Clustering: An Empirical Study of Triplet Loss Approaches

Abstract
In this work, we evaluate two different image clustering objectives, k-means clustering and correlation clustering, in the context of Triplet Loss induced feature space embeddings. Specifically, we train a convolutional neural network to learn discriminative features by optimizing two popular versions of the Triplet Loss in order to study their clustering properties under the assumption of noisy labels. Additionally, we propose a new, simple Triplet Loss formulation, which shows desirable properties with respect to formal clustering objectives and outperforms the existing methods. We evaluate all three Triplet loss formulations for K-means and correlation clustering on the CIFAR-10 image classification dataset.
Author(s)
Ho, Kalun
Fraunhofer-Institut für Techno- und Wirtschaftsmathematik ITWM
Keuper, Janis
Institute for Machine Learning and Analytics (IMLA), Offenburg University
Pfreundt, Franz-Josef
Fraunhofer-Institut für Techno- und Wirtschaftsmathematik ITWM
Keuper, Margret
Data and Web Science Group, University of Mannheim
Hauptwerk
ICPR 2020, 25th International Conference on Pattern Recognition. Proceedings
Funder
Deutsche Forschungsgemeinschaft DFG
Konferenz
International Conference on Pattern Recognition (ICPR) 2021
Thumbnail Image
DOI
10.1109/ICPR48806.2021.9412602
Language
English
google-scholar
Fraunhofer-Institut für Techno- und Wirtschaftsmathematik ITWM
  • Cookie settings
  • Imprint
  • Privacy policy
  • Api
  • Send Feedback
© 2022