Skip to content

Research at St Andrews

Texture features for object salience

Research output: Contribution to journalArticlepeer-review

Author(s)

Kasim Terzić, Sai Krishna, J. M. H. du Buf

School/Research organisations

Abstract

Although texture is important for many vision-related tasks, it is not used in most salience models. As a consequence, there are images where all existing salience algorithms fail. We introduce a novel set of texture features built on top of a fast model of complex cells in striate cortex, i.e., visual area V1. The texture at each position is characterised by the two-dimensional local power spectrum obtained from Gabor filters which are tuned to many scales and orientations. We then apply a parametric model and describe the local spectrum by the combination of two one-dimensional Gaussian approximations: the scale and orientation distributions. The scale distribution indicates whether the texture has a dominant frequency and what frequency it is. Likewise, the orientation distribution attests the degree of anisotropy. We evaluate the features in combination with the state-of-the-art VOCUS2 salience algorithm. We found that using our novel texture features in addition to colour improves AUC by 3.8% on the PASCAL-S dataset when compared to the colour-only baseline, and by 62% on a novel texture-based dataset.
Close

Details

Original languageEnglish
Pages (from-to)43-51
JournalImage and Vision Computing
Volume67
Early online date22 Sep 2017
DOIs
Publication statusPublished - Nov 2017

    Research areas

  • Texture, Colour, Salience, Attention, Benchmark

Discover related content
Find related publications, people, projects and more using interactive charts.

View graph of relations

Related by author

  1. Few-shot linguistic grounding of visual attributes and relations using gaussian kernels

    Koudouna, D. & Terzić, K., 8 Feb 2021, Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - (Volume 5). Farinella, G. M., Radeva, P., Braz, J. & Bouatouch, K. (eds.). SCITEPRESS - Science and Technology Publications, Vol. 5 VISAPP. p. 146-156

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

  2. Visualization as Intermediate Representations (VLAIR) for human activity recognition

    Jiang, A., Nacenta, M., Terzić, K. & Ye, J., 18 May 2020, PervasiveHealth '20: Proceedings of the 14th EAI International Conference on Pervasive Computing Technologies for Healthcare. Munson, S. A. & Schueller, S. M. (eds.). ACM, p. 201-210 10 p.

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

  3. Supervisor recommendation tool for Computer Science projects

    Zemaityte, G. & Terzic, K., 9 Jan 2019, Proceedings of the 3rd Conference on Computing Education Practice (CEP '19) . New York: ACM, 4 p. 1

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

  4. BINK: Biological Binary Keypoint Descriptor

    Saleiro, M., Terzić, K., Rodrigues, J. M. F. & du Buf, J. M. H., Dec 2017, In: BioSystems. 162, p. 147-156

    Research output: Contribution to journalArticlepeer-review

  5. Interpretable feature maps for robot attention

    Terzić, K. & du Buf, J. M. H., 2017, Universal Access in Human–Computer Interaction. Design and Development Approaches and Methods: 11th International Conference, UAHCI 2017, Held as Part of HCI International 2017, Vancouver, BC, Canada, July 9–14, 2017, Proceedings, Part I. Antona, M. & Stephanidis, C. (eds.). Cham: Springer, p. 456-467 12 p. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); vol. 10277).

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

Related by journal

  1. Digitization of non-regular shapes in arbitrary dimensions

    Stelldinger, P. & Terzic, K., 1 Oct 2008, In: Image and Vision Computing. 26, 10, p. 1338-1346 9 p.

    Research output: Contribution to journalArticlepeer-review

  2. An information-theoretic approach to face recognition from face motion manifolds

    Arandelovic, O. & Cipolla, R., 1 Jun 2006, In: Image and Vision Computing. 24, 6, p. 639-647 9 p.

    Research output: Contribution to journalArticlepeer-review

ID: 251156304

Top