English  Sprachen Icon  |  Gebärdensprache  |  Leichte Sprache  |  Kontakt


Humans search all the time. We want to know how.

Von Kilian Semmelmann (17.12.2013)

Humans search all the time. They search for keys in kitchens, for bathroom signs in restaurants, for their kids in a horde of first graders or for special offers in their newspapers. This field of looking at how you look for something is called “Visual Search”, a subfield of “Visual Attention”. And it has important real life applications: Two examples of professional searchers would be the security officers at airports who scan luggage to identify potential dangerous items or radiologists in hospitals, who check x-ray or CT scans for abnormalities. But before going on, I want you to get a better idea of what we do by participating in our current online study about this topic: 


It will only take about 30 Minutes and your task will be to search for gas stations – we will explain why this is important after the task. Please do this before continuing reading; otherwise you will be biased through the knowledge provided below.

Semmelmann: Fig. 1[Bildunterschrift / Subline]: Figure 1. Simplified security scan of carry-on luggage. How can we differentiate dangerous from harmless items?
[Bildunterschrift / Subline]: Figure 2. Mapforaging: How many gas stations can you find?

Done? Very good, thank you. As you can imagine, most research in the field of Visual Attention concentrates on what attracts your attention and how we can influence these processes. Our project on the other hand tries to investigate the processes at the very end of the search process: When do we decide, that there are no (more) targets in the area we are currently searching? When and why do we decide to switch from one display to another? What happens if we search on empty displays? That is exactly what is done in our research study mentioned above: Your task is to search for gas stations on aerial maps and mark them. This is realized through a web application using Google Maps and allows you to zoom and navigate the map as you like. Additionally, you are free to change the current display to a new one, whenever you have the feeling that it would be more efficient to change to another display. 

Luckily, similar questions are asked in a different field of science, namely the field of Behavioral Ecology, especially in Animal Foraging Behavior. Here the situation is as follows: An animal (e.g. a bird) is foraging at a bush (one ‘food’-patch) and decides at a specific point of time to switch to a different patch. Normally it would not deplete the whole bush, but rather behaves in an optimal way (Charnov’s Marginal Value Theorem) by changing to a fresh patch as soon as the income (berries) at the current patch drops below the average income of all patches in the environment. That means: If it is more efficient to take the cost of travelling to the next bush instead of staying at the current, it will change.

[Bildunterschrift / Subline]: Figure 3. Preliminary result of a small sample study: Participants switch displays as soon as the income rate (blue line) drops below average income (red line) rate. It shows that humans might behave optimal in the search task.
Semmelmann: Fig. 4[Bildunterschrift / Subline]: Figure 4. Three possible search strategies: strategic (use of underlying information), sample (covering map through sinusoidal movement) and random searching.

Our hypothesis is that human visual search behavior as described above might behave in a similar way as animal foraging behavior. Coming back to our online experiment, it means that you switch to the next map, as soon as it gets too exhaustive to find another possible station on your current map. A first small-scale study revealed that this might be the case: As soon as the income-rate of gas stations per time drops below the average, participants switch to the next map (figure 3), how animals would travel to the next bush. That means they intrinsically behave in an optimal way when searching for targets. Moreover, we have been able to identify several unique search strategies that people use in this task: Strategic (using underlying information), sampling (using a lawn-mower movement to cover the whole map) or random searching (figure 4). To further these first findings, we need way more data for statistical analysis, what led to publishing the study as an online version for everyone to participate.

To sum up, we might want to recall why it is important to know how people behave when searching for objects. Remember the security officers and radiologists, who decide every day whether a display (suitcase or x-ray) is safe. Being under time pressure, they have to decide whether they can quit searching or not. If they would fail this task, the consequences would be immense. Thus, our research aims to identify and possibly modify the factors how humans search and when they quit their search to improve search efficiency and allow for a safer, more reliable environment. And you could help by participating in our study and distributing the link – thank you very much. 

Wissenschaftlicher Werdegang
  • 2008-2011
  • B.Sc. in Cognitive Science, University of Osnabrück
  • 2011-2013
  • M.Sc. in Neuro-cognitve Psychology, Ludwig-Maximilians-University Munich

Scholarships and Awards
  • * DAAD: Research Internships in Science and Engineering (2010)
  • * LMU PROMOS + DAAD PROSA scholarships (2012 and 2013)
  • * "Leader of Tomorrow", St. Gallen Wings of Excellence Award (2013)