Interpretable feature maps for robot attention

Kasim Terzić; J. M.H. du Buf

doi:10.1007/978-3-319-58706-6_37

Interpretable feature maps for robot attention

Kasim Terzić^*, J. M.H. du Buf

^*Corresponding author for this work

School of Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Attention is crucial for autonomous agents interacting with complex environments. In a real scenario, our expectations drive attention, as we look for crucial objects to complete our understanding of the scene. But most visual attention models to date are designed to drive attention in a bottom-up fashion, without context, and the features they use are not always suitable for driving top-down attention. In this paper, we present an attentional mechanism based on semantically meaningful, interpretable features. We show how to generate a low-level semantic representation of the scene in real time, which can be used to search for objects based on specific features such as colour, shape, orientation, speed, and texture.

Original language	English
Title of host publication	Universal Access in Human–Computer Interaction. Design and Development Approaches and Methods
Subtitle of host publication	11th International Conference, UAHCI 2017, Held as Part of HCI International 2017, Vancouver, BC, Canada, July 9–14, 2017, Proceedings, Part I
Editors	Margherita Antona, Constantine Stephanidis
Place of Publication	Cham
Publisher	Springer
Pages	456-467
Number of pages	12
ISBN (Electronic)	9783319587066
ISBN (Print)	9783319587059
DOIs	https://doi.org/10.1007/978-3-319-58706-6_37
Publication status	Published - 2017
Event	11th International Conference on Universal Access in Human-Computer Interaction, UAHCI 2017, held as part of the 19th International Conference on Human-Computer Interaction, HCI 2017 - Vancouver, Canada Duration: 9 Jul 2017 → 14 Jul 2017 Conference number: 11 http://2017.hci.international/index.php

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10277
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	11th International Conference on Universal Access in Human-Computer Interaction, UAHCI 2017, held as part of the 19th International Conference on Human-Computer Interaction, HCI 2017
Abbreviated title	UAHCI
Country/Territory	Canada
City	Vancouver
Period	9/07/17 → 14/07/17
Internet address	http://2017.hci.international/index.php

Access to Document

10.1007/978-3-319-58706-6_37

Terzic-2017-Interpretable-feature-UAHCI-AAM
© 2017, Springer. This work has been made available online in accordance with the publisher’s policies. This is the author created accepted version manuscript following peer review and as such may differ slightly from the final published version. The final published version of this work is available at https://doi.org/10.1007/978-3-319-58706-6_37
Accepted author manuscript, 891 KB

Cite this

Terzić, K., & du Buf, J. M. H. (2017). Interpretable feature maps for robot attention. In M. Antona, & C. Stephanidis (Eds.), 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 (pp. 456-467). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10277). Springer. https://doi.org/10.1007/978-3-319-58706-6_37

Terzić, Kasim ; du Buf, J. M.H. / Interpretable feature maps for robot attention. 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. editor / Margherita Antona ; Constantine Stephanidis. Cham : Springer, 2017. pp. 456-467 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{725a4c9a2ad1491780a49232d5ff1ccc,

title = "Interpretable feature maps for robot attention",

abstract = "Attention is crucial for autonomous agents interacting with complex environments. In a real scenario, our expectations drive attention, as we look for crucial objects to complete our understanding of the scene. But most visual attention models to date are designed to drive attention in a bottom-up fashion, without context, and the features they use are not always suitable for driving top-down attention. In this paper, we present an attentional mechanism based on semantically meaningful, interpretable features. We show how to generate a low-level semantic representation of the scene in real time, which can be used to search for objects based on specific features such as colour, shape, orientation, speed, and texture.",

author = "Kasim Terzi{\'c} and {du Buf}, {J. M.H.}",

year = "2017",

doi = "10.1007/978-3-319-58706-6_37",

language = "English",

isbn = "9783319587059",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "456--467",

editor = "Margherita Antona and Constantine Stephanidis",

booktitle = "Universal Access in Human–Computer Interaction. Design and Development Approaches and Methods",

address = "Netherlands",

note = "11th International Conference on Universal Access in Human-Computer Interaction, UAHCI 2017, held as part of the 19th International Conference on Human-Computer Interaction, HCI 2017, UAHCI ; Conference date: 09-07-2017 Through 14-07-2017",

url = "http://2017.hci.international/index.php",

}

Terzić, K & du Buf, JMH 2017, Interpretable feature maps for robot attention. in M Antona & C Stephanidis (eds), 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. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10277, Springer, Cham, pp. 456-467, 11th International Conference on Universal Access in Human-Computer Interaction, UAHCI 2017, held as part of the 19th International Conference on Human-Computer Interaction, HCI 2017, Vancouver, British Columbia, Canada, 9/07/17. https://doi.org/10.1007/978-3-319-58706-6_37

Interpretable feature maps for robot attention. / Terzić, Kasim; du Buf, J. M.H.
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. ed. / Margherita Antona; Constantine Stephanidis. Cham: Springer, 2017. p. 456-467 (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 proceeding › Conference contribution

TY - GEN

T1 - Interpretable feature maps for robot attention

AU - Terzić, Kasim

AU - du Buf, J. M.H.

N1 - Conference code: 11

PY - 2017

Y1 - 2017

N2 - Attention is crucial for autonomous agents interacting with complex environments. In a real scenario, our expectations drive attention, as we look for crucial objects to complete our understanding of the scene. But most visual attention models to date are designed to drive attention in a bottom-up fashion, without context, and the features they use are not always suitable for driving top-down attention. In this paper, we present an attentional mechanism based on semantically meaningful, interpretable features. We show how to generate a low-level semantic representation of the scene in real time, which can be used to search for objects based on specific features such as colour, shape, orientation, speed, and texture.

AB - Attention is crucial for autonomous agents interacting with complex environments. In a real scenario, our expectations drive attention, as we look for crucial objects to complete our understanding of the scene. But most visual attention models to date are designed to drive attention in a bottom-up fashion, without context, and the features they use are not always suitable for driving top-down attention. In this paper, we present an attentional mechanism based on semantically meaningful, interpretable features. We show how to generate a low-level semantic representation of the scene in real time, which can be used to search for objects based on specific features such as colour, shape, orientation, speed, and texture.

U2 - 10.1007/978-3-319-58706-6_37

DO - 10.1007/978-3-319-58706-6_37

M3 - Conference contribution

AN - SCOPUS:85025168961

SN - 9783319587059

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 456

EP - 467

BT - Universal Access in Human–Computer Interaction. Design and Development Approaches and Methods

A2 - Antona, Margherita

A2 - Stephanidis, Constantine

PB - Springer

CY - Cham

T2 - 11th International Conference on Universal Access in Human-Computer Interaction, UAHCI 2017, held as part of the 19th International Conference on Human-Computer Interaction, HCI 2017

Y2 - 9 July 2017 through 14 July 2017

ER -

Terzić K, du Buf JMH. Interpretable feature maps for robot attention. In Antona M, Stephanidis C, editors, 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. Cham: Springer. 2017. p. 456-467. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). Epub 2017 May 16. doi: 10.1007/978-3-319-58706-6_37

Interpretable feature maps for robot attention

Abstract

Publication series

Conference

Access to Document

Fingerprint

Cite this