The joint submission of the TU Berlin and Fraunhofer FIRST (TUBFI) to the image CLEF 2011 photo annotation task

Alexander Binder, Wojciech Samek, Marius Kloft, Christina Müller, Klaus Muller, Motoaki Kawanabe

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

11 Citations (Scopus)

Abstract

In this paper we present details on the joint submission of TU Berlin and Fraunhofer FIRST to the ImageCLEF 2011 Photo Annotation Task.We sought to experiment with extensions of Bag-of-Words (BoW) models at several levels and to apply several kernel-based learning methods recently developed in our group. For classifier training we used non-sparse multiple kernel learning (MKL) and an efficient multi-task learning (MTL) heuristic based on MKL over kernels from classifier outputs. For the multi-modal fusion we used a smoothing method on tag-based features inspired by Bag-of-Words soft mappings and Markov random walks. We submitted one multi-modal run extended by the user tags and four purely visual runs based on Bag-of-Words models. Our best visual result which used the MTL method was ranked first according to mean average precision (MAP) within the purely visual submissions. Our multi-modal submission achieved the first rank by MAP among the multi-modal submissions and the best MAP among all submissions. Submissions by other groups such as BPACAD, CAEN, UvA-ISIS, LIRIS were ranked closely.

Original languageEnglish
Title of host publicationCEUR Workshop Proceedings
PublisherCEUR-WS
Volume1177
Publication statusPublished - 2011
Externally publishedYes
Event2011 Working Notes for CLEF Conference, CLEF 2011 - Amsterdam, Netherlands
Duration: 2011 Sep 192011 Sep 22

Other

Other2011 Working Notes for CLEF Conference, CLEF 2011
CountryNetherlands
CityAmsterdam
Period11/9/1911/9/22

Fingerprint

Classifiers
Fusion reactions
Experiments

Keywords

  • Bag-of-words
  • Image classification
  • Image clef
  • Multi-task learning
  • Multiple kernel learning
  • Photo annotation
  • Theseus

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Binder, A., Samek, W., Kloft, M., Müller, C., Muller, K., & Kawanabe, M. (2011). The joint submission of the TU Berlin and Fraunhofer FIRST (TUBFI) to the image CLEF 2011 photo annotation task. In CEUR Workshop Proceedings (Vol. 1177). CEUR-WS.

The joint submission of the TU Berlin and Fraunhofer FIRST (TUBFI) to the image CLEF 2011 photo annotation task. / Binder, Alexander; Samek, Wojciech; Kloft, Marius; Müller, Christina; Muller, Klaus; Kawanabe, Motoaki.

CEUR Workshop Proceedings. Vol. 1177 CEUR-WS, 2011.

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

Binder, A, Samek, W, Kloft, M, Müller, C, Muller, K & Kawanabe, M 2011, The joint submission of the TU Berlin and Fraunhofer FIRST (TUBFI) to the image CLEF 2011 photo annotation task. in CEUR Workshop Proceedings. vol. 1177, CEUR-WS, 2011 Working Notes for CLEF Conference, CLEF 2011, Amsterdam, Netherlands, 11/9/19.
Binder A, Samek W, Kloft M, Müller C, Muller K, Kawanabe M. The joint submission of the TU Berlin and Fraunhofer FIRST (TUBFI) to the image CLEF 2011 photo annotation task. In CEUR Workshop Proceedings. Vol. 1177. CEUR-WS. 2011
Binder, Alexander ; Samek, Wojciech ; Kloft, Marius ; Müller, Christina ; Muller, Klaus ; Kawanabe, Motoaki. / The joint submission of the TU Berlin and Fraunhofer FIRST (TUBFI) to the image CLEF 2011 photo annotation task. CEUR Workshop Proceedings. Vol. 1177 CEUR-WS, 2011.
@inproceedings{b1b7ff0e2f6c4570beddbf38605e4f81,
title = "The joint submission of the TU Berlin and Fraunhofer FIRST (TUBFI) to the image CLEF 2011 photo annotation task",
abstract = "In this paper we present details on the joint submission of TU Berlin and Fraunhofer FIRST to the ImageCLEF 2011 Photo Annotation Task.We sought to experiment with extensions of Bag-of-Words (BoW) models at several levels and to apply several kernel-based learning methods recently developed in our group. For classifier training we used non-sparse multiple kernel learning (MKL) and an efficient multi-task learning (MTL) heuristic based on MKL over kernels from classifier outputs. For the multi-modal fusion we used a smoothing method on tag-based features inspired by Bag-of-Words soft mappings and Markov random walks. We submitted one multi-modal run extended by the user tags and four purely visual runs based on Bag-of-Words models. Our best visual result which used the MTL method was ranked first according to mean average precision (MAP) within the purely visual submissions. Our multi-modal submission achieved the first rank by MAP among the multi-modal submissions and the best MAP among all submissions. Submissions by other groups such as BPACAD, CAEN, UvA-ISIS, LIRIS were ranked closely.",
keywords = "Bag-of-words, Image classification, Image clef, Multi-task learning, Multiple kernel learning, Photo annotation, Theseus",
author = "Alexander Binder and Wojciech Samek and Marius Kloft and Christina M{\"u}ller and Klaus Muller and Motoaki Kawanabe",
year = "2011",
language = "English",
volume = "1177",
booktitle = "CEUR Workshop Proceedings",
publisher = "CEUR-WS",

}

TY - GEN

T1 - The joint submission of the TU Berlin and Fraunhofer FIRST (TUBFI) to the image CLEF 2011 photo annotation task

AU - Binder, Alexander

AU - Samek, Wojciech

AU - Kloft, Marius

AU - Müller, Christina

AU - Muller, Klaus

AU - Kawanabe, Motoaki

PY - 2011

Y1 - 2011

N2 - In this paper we present details on the joint submission of TU Berlin and Fraunhofer FIRST to the ImageCLEF 2011 Photo Annotation Task.We sought to experiment with extensions of Bag-of-Words (BoW) models at several levels and to apply several kernel-based learning methods recently developed in our group. For classifier training we used non-sparse multiple kernel learning (MKL) and an efficient multi-task learning (MTL) heuristic based on MKL over kernels from classifier outputs. For the multi-modal fusion we used a smoothing method on tag-based features inspired by Bag-of-Words soft mappings and Markov random walks. We submitted one multi-modal run extended by the user tags and four purely visual runs based on Bag-of-Words models. Our best visual result which used the MTL method was ranked first according to mean average precision (MAP) within the purely visual submissions. Our multi-modal submission achieved the first rank by MAP among the multi-modal submissions and the best MAP among all submissions. Submissions by other groups such as BPACAD, CAEN, UvA-ISIS, LIRIS were ranked closely.

AB - In this paper we present details on the joint submission of TU Berlin and Fraunhofer FIRST to the ImageCLEF 2011 Photo Annotation Task.We sought to experiment with extensions of Bag-of-Words (BoW) models at several levels and to apply several kernel-based learning methods recently developed in our group. For classifier training we used non-sparse multiple kernel learning (MKL) and an efficient multi-task learning (MTL) heuristic based on MKL over kernels from classifier outputs. For the multi-modal fusion we used a smoothing method on tag-based features inspired by Bag-of-Words soft mappings and Markov random walks. We submitted one multi-modal run extended by the user tags and four purely visual runs based on Bag-of-Words models. Our best visual result which used the MTL method was ranked first according to mean average precision (MAP) within the purely visual submissions. Our multi-modal submission achieved the first rank by MAP among the multi-modal submissions and the best MAP among all submissions. Submissions by other groups such as BPACAD, CAEN, UvA-ISIS, LIRIS were ranked closely.

KW - Bag-of-words

KW - Image classification

KW - Image clef

KW - Multi-task learning

KW - Multiple kernel learning

KW - Photo annotation

KW - Theseus

UR - http://www.scopus.com/inward/record.url?scp=84922032448&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84922032448&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84922032448

VL - 1177

BT - CEUR Workshop Proceedings

PB - CEUR-WS

ER -