Behavioral biases when viewing multiplexed scenes

Scene structure and frames of reference for inspection

Matthew J. Stainer, Kenneth C. Scott-Brown, Benjamin William Tatler*

*Corresponding author for this work

Research output: Contribution to journalArticle

13 Citations (Scopus)
4 Downloads (Pure)

Abstract

Where people look when viewing a scene has been a much explored avenue of vision research (e.g., see Tatler, 2009). Current understanding of eye guidance suggests that a combination of high and low-level factors influence fixation selection (e.g., Torralba et al., 2006), but that there are also strong biases toward the center of an image (Tatler, 2007). However, situations where we view multiplexed scenes are becoming increasingly common, and it is unclear how visual inspection might be arranged when content lacks normal semantic or spatial structure. Here we use the central bias to examine how gaze behavior is organized in scenes that are presented in their normal format, or disrupted by scrambling the quadrants and separating them by space. In Experiment 1, scrambling scenes had the strongest influence on gaze allocation. Observers were highly biased by the quadrant center, although physical space did not enhance this bias. However, the center of the display still contributed to fixation selection above chance, and was most influential early in scene viewing. When the top left quadrant was held constant across all conditions in Experiment 2, fixation behavior was significantly influenced by the overall arrangement of the display, with fixations being biased toward the quadrant center when the other three quadrants were scrambled (despite the visual information in this quadrant being identical in all conditions). When scenes are scrambled into four quadrants and semantic contiguity is disrupted, observers no longer appear to view the content as a single scene (despite it consisting of the same visual information overall), but rather anchor visual inspection around the four separate "sub-scenes." Moreover, the frame of reference that observers use when viewing the multiplex seems to change across viewing time: from an early bias toward the display center to a later bias toward quadrant centers.

Original languageEnglish
Article number624
Pages (from-to)1-12
Number of pages12
JournalFrontiers in Psychology
Volume4
DOIs
Publication statusPublished - 24 Sep 2013

Fingerprint

Semantics
Research

Keywords

  • Central bias
  • Frames of reference
  • Multiplex
  • Scene structure
  • Scene viewing

ASJC Scopus subject areas

  • Psychology(all)

Cite this

Behavioral biases when viewing multiplexed scenes : Scene structure and frames of reference for inspection. / Stainer, Matthew J.; Scott-Brown, Kenneth C.; Tatler, Benjamin William.

In: Frontiers in Psychology, Vol. 4, 624, 24.09.2013, p. 1-12.

Research output: Contribution to journalArticle

@article{91a3789ac2d84406aa3d0faeeab3907c,
title = "Behavioral biases when viewing multiplexed scenes: Scene structure and frames of reference for inspection",
abstract = "Where people look when viewing a scene has been a much explored avenue of vision research (e.g., see Tatler, 2009). Current understanding of eye guidance suggests that a combination of high and low-level factors influence fixation selection (e.g., Torralba et al., 2006), but that there are also strong biases toward the center of an image (Tatler, 2007). However, situations where we view multiplexed scenes are becoming increasingly common, and it is unclear how visual inspection might be arranged when content lacks normal semantic or spatial structure. Here we use the central bias to examine how gaze behavior is organized in scenes that are presented in their normal format, or disrupted by scrambling the quadrants and separating them by space. In Experiment 1, scrambling scenes had the strongest influence on gaze allocation. Observers were highly biased by the quadrant center, although physical space did not enhance this bias. However, the center of the display still contributed to fixation selection above chance, and was most influential early in scene viewing. When the top left quadrant was held constant across all conditions in Experiment 2, fixation behavior was significantly influenced by the overall arrangement of the display, with fixations being biased toward the quadrant center when the other three quadrants were scrambled (despite the visual information in this quadrant being identical in all conditions). When scenes are scrambled into four quadrants and semantic contiguity is disrupted, observers no longer appear to view the content as a single scene (despite it consisting of the same visual information overall), but rather anchor visual inspection around the four separate {"}sub-scenes.{"} Moreover, the frame of reference that observers use when viewing the multiplex seems to change across viewing time: from an early bias toward the display center to a later bias toward quadrant centers.",
keywords = "Central bias, Frames of reference, Multiplex, Scene structure, Scene viewing",
author = "Stainer, {Matthew J.} and Scott-Brown, {Kenneth C.} and Tatler, {Benjamin William}",
note = "The authors would like to thank Yoriko Hirose for assistance with data collection.",
year = "2013",
month = "9",
day = "24",
doi = "10.3389/fpsyg.2013.00624",
language = "English",
volume = "4",
pages = "1--12",
journal = "Frontiers in Psychology",
issn = "1664-1078",
publisher = "Frontiers Media S.A.",

}

TY - JOUR

T1 - Behavioral biases when viewing multiplexed scenes

T2 - Scene structure and frames of reference for inspection

AU - Stainer, Matthew J.

AU - Scott-Brown, Kenneth C.

AU - Tatler, Benjamin William

N1 - The authors would like to thank Yoriko Hirose for assistance with data collection.

PY - 2013/9/24

Y1 - 2013/9/24

N2 - Where people look when viewing a scene has been a much explored avenue of vision research (e.g., see Tatler, 2009). Current understanding of eye guidance suggests that a combination of high and low-level factors influence fixation selection (e.g., Torralba et al., 2006), but that there are also strong biases toward the center of an image (Tatler, 2007). However, situations where we view multiplexed scenes are becoming increasingly common, and it is unclear how visual inspection might be arranged when content lacks normal semantic or spatial structure. Here we use the central bias to examine how gaze behavior is organized in scenes that are presented in their normal format, or disrupted by scrambling the quadrants and separating them by space. In Experiment 1, scrambling scenes had the strongest influence on gaze allocation. Observers were highly biased by the quadrant center, although physical space did not enhance this bias. However, the center of the display still contributed to fixation selection above chance, and was most influential early in scene viewing. When the top left quadrant was held constant across all conditions in Experiment 2, fixation behavior was significantly influenced by the overall arrangement of the display, with fixations being biased toward the quadrant center when the other three quadrants were scrambled (despite the visual information in this quadrant being identical in all conditions). When scenes are scrambled into four quadrants and semantic contiguity is disrupted, observers no longer appear to view the content as a single scene (despite it consisting of the same visual information overall), but rather anchor visual inspection around the four separate "sub-scenes." Moreover, the frame of reference that observers use when viewing the multiplex seems to change across viewing time: from an early bias toward the display center to a later bias toward quadrant centers.

AB - Where people look when viewing a scene has been a much explored avenue of vision research (e.g., see Tatler, 2009). Current understanding of eye guidance suggests that a combination of high and low-level factors influence fixation selection (e.g., Torralba et al., 2006), but that there are also strong biases toward the center of an image (Tatler, 2007). However, situations where we view multiplexed scenes are becoming increasingly common, and it is unclear how visual inspection might be arranged when content lacks normal semantic or spatial structure. Here we use the central bias to examine how gaze behavior is organized in scenes that are presented in their normal format, or disrupted by scrambling the quadrants and separating them by space. In Experiment 1, scrambling scenes had the strongest influence on gaze allocation. Observers were highly biased by the quadrant center, although physical space did not enhance this bias. However, the center of the display still contributed to fixation selection above chance, and was most influential early in scene viewing. When the top left quadrant was held constant across all conditions in Experiment 2, fixation behavior was significantly influenced by the overall arrangement of the display, with fixations being biased toward the quadrant center when the other three quadrants were scrambled (despite the visual information in this quadrant being identical in all conditions). When scenes are scrambled into four quadrants and semantic contiguity is disrupted, observers no longer appear to view the content as a single scene (despite it consisting of the same visual information overall), but rather anchor visual inspection around the four separate "sub-scenes." Moreover, the frame of reference that observers use when viewing the multiplex seems to change across viewing time: from an early bias toward the display center to a later bias toward quadrant centers.

KW - Central bias

KW - Frames of reference

KW - Multiplex

KW - Scene structure

KW - Scene viewing

UR - http://www.scopus.com/inward/record.url?scp=84885332351&partnerID=8YFLogxK

U2 - 10.3389/fpsyg.2013.00624

DO - 10.3389/fpsyg.2013.00624

M3 - Article

VL - 4

SP - 1

EP - 12

JO - Frontiers in Psychology

JF - Frontiers in Psychology

SN - 1664-1078

M1 - 624

ER -