Case Study: Spot the Dog Within 120 Milliseconds

Background image

Researchers at the French National Center for Scientific Research (CNRS) used SMI high speed eye tracking to investigate the hierarchy of object categorization with a “forced-choice-saccade” paradigm.

Challenge

How long does it take someone to recognize an object in a scene? As the title of a publication in the Journal of Cognitive Neuroscience makes clear: "At 120 ms you can spot the animal but you don't yet know it's a dog". This study examines how the visual system processes information at a basic level (dogs), subordinate level (Dalmatian) and a superordinate level (animals) and asks "what can we perceive with just a glance at a scene?"

Solution

The study involved volunteers using an SMI high-speed eye tracker to view photographic images flashed up on screen. While a target and a distractor image were displayed simultaneously, the eye tracker measured the very fast movements of the eye as it "saccades to" the relevant image – a "forced-choice-saccade task".

The experiments varied in the level of categorization and the morphological similarity within the sets of targets and distractors. For example, some participants were shown animals and non-animal distractor images such as a building, a wooden log or even a photo of a bathroom. In other cases, both images were of a mammal, such as a dog and cat image side by side.

SMI Eye Tracking

High-speed eye tracking from SMI was critical to this study, given that the researchers were gauging the extremely fast eye movements that follow a forced-choice saccadic task to measure visual processing speed.

Benefit

The study found that participants were very good at making saccades towards images of animals when the other option was a non-animal image. Accuracy was high (around 80% correct) for categorization tasks that can be solved at this superordinate level.

But critically, the study found, "accuracy fell when participants were required to saccade toward images among distractors that contained other animals". Accuracy dropped to almost chance levels for basic level categorization.

Researchers concluded that information about the superordinate category is available earlier during visual object processing. It seems that the visual system takes more time to recognize objects at the basic level than at the superordinate level.

Background

The National Center for Scientific Research, or CNRS, is a public organization under the responsibility of the French Ministry of Education and Research.

Benefit

The study found that participants were very good at making saccades towards images of animals when the other option was a non-animal image. Accuracy was high (around 80% correct) for categorization tasks that can be solved at this superordinate level.

But critically, the study found, "accuracy fell when participants were required to saccade toward images among distractors that contained other animals". Accuracy dropped to almost chance levels for basic level categorization.

Researchers concluded that information about the superordinate category is available earlier during visual object processing. It seems that the visual system takes more time to recognize objects at the basic level than at the superordinate level.

Customer Institution French National Center for Scientific Research (CNRS)
Customer website http://www.cnrs.fr/index.php
Related publication