Insights from classifying visual concepts with multiple kernel learning

Research output: Journal contributionsJournal articlesResearchpeer-review

Authors

  • Alexander Binder
  • Shinichi Nakajima
  • Marius Kloft
  • Christina Müller
  • Wojciech Samek
  • Ulf Brefeld
  • Klaus Robert Müller
  • Motoaki Kawanabe

Combining information from various image features has become a standard technique in concept recognition tasks. However, the optimal way of fusing the resulting kernel functions is usually unknown in practical applications. Multiple kernel learning (MKL) techniques allow to determine an optimal linear combination of such similarity matrices. Classical approaches to MKL promote sparse mixtures. Unfortunately, 1-norm regularized MKL variants are often observed to be outperformed by an unweighted sum kernel. The main contributions of this paper are the following: we apply a recently developed non-sparse MKL variant to state-of-the-art concept recognition tasks from the application domain of computer vision. We provide insights on benefits and limits of non-sparse MKL and compare it against its direct competitors, the sum-kernel SVM and sparse MKL. We report empirical results for the PASCAL VOC 2009 Classification and ImageCLEF2010 Photo Annotation challenge data sets. Data sets (kernel matrices) as well as further information are available at http://doc.ml.tuberlin.de/image_mkl/(Accessed 2012 Jun 25).

Original languageEnglish
Article numbere38897
JournalPLoS ONE
Volume7
Issue number8
ISSN1932-6203
DOIs
Publication statusPublished - 24.08.2012
Externally publishedYes

    Research areas

  • Informatics - concept formation, controlled study, histogram, image display, intermethod comparison, kernel method, machine learning, scoring system, support vector machine, task performance, validity
  • Business informatics

Documents

DOI

Recently viewed

Publications

  1. Interpreting Strings, Weaving Threads
  2. Proceedings of the SeMantic Answer Type and Relation Prediction Task at ISWC 2021 Semantic Web Challenge (SMART2021)
  3. N-term approximation in anisotropic function spaces
  4. Fostering Circularity: Building a Local Community and Implementing Circular Processes
  5. Comparing Two Voltage Observers in a Sensorsystem using Repetitive Control
  6. Using haar wavelets for fault detection in technical processes
  7. Multi-view learning with dependent views
  8. Making an Impression Through Openness
  9. Selecting and Adapting Methods for Analysis and Design in Value-Sensitive Digital Social Innovation Projects: Toward Design Principles
  10. Problem structuring for transitions
  11. On the Functional Controllability Using a Geometric Approach together with a Decoupled MPC for Motion Control in Robotino
  12. Age effects on controlling tools with sensorimotor transformations
  13. A guided simulated annealing search for solving the pick-up and delivery problem with time windows and capacity constraints
  14. Efficient and accurate ℓ p-norm multiple kernel learning
  15. Partitioned beta diversity patterns of plants across sharp and distinct boundaries of quartz habitat islands
  16. Reading and Calculating in Word Problem Solving
  17. Intentionality
  18. Taking notes as a strategy for solving reality-based tasks in mathematics
  19. Assessing Effects Through Semi-Field and Field Toxicity Testing
  20. Linear free vibrations with uncertain initial conditions
  21. Some model properties to control a permanent magnet machine using a controlled invariant subspace
  22. Robust Control of Mobile Transportation Object with 3D Technical Vision System
  23. An Interactive Layers Model of Self-Regulated Learning and Cognitive Load
  24. Speed of processing and stimulus complexity in low-frequency and high-frequency channels
  25. Support vector machines with example dependent costs
  26. Mathematical Modeling for Robot 3D Laser Scanning in Complete Darkness Environments to Advance Pipeline Inspection
  27. Trajectory tracking using MPC and a velocity observer for flat actuator systems in automotive applications
  28. Interactions between ecosystem properties and land use clarify spatial strategies to optimize trade-offs between agriculture and species conservation
  29. Multiphase-field modeling of temperature-driven intermetallic compound evolution in an Al-Mg system for application to solid-state joining processes
  30. HAWK - hybrid question answering using linked data
  31. Erroneous examples as desirable difficulty
  32. Student Game Design for Language Learning
  33. The identification of up-And downstream industries using input-output tables and a firm-level application to minority shareholdings
  34. Automatic three-dimensional geometry and mesh generation of periodic representative volume elements for matrix-inclusion composites