Course Home
Administration
Syllabus
References
Lecture Notes
Readings
Assignments
Projects
Resources


Comments? Idea?























































































































































































































































































































































Which objects we notice first?

Final project, by

Ruti Shamay and Inbar Mizrahi

rutishamay@gmail.com

mizrahi.inbar@gmail.com


Introduction

In this report, we’ll present an experiment that we maid. The experiment was to examine to which objects we notice first. There were 20 participants, and 50 pictures were presented to each of them randomly. Each picture was presented to a random time, to one from the following options: 35, 100, 300 and 980 milliseconds. After each picture there was presented the same mask to prevent the “burn” of the information, and then the participant was ask to write which objects he saw in the picture in many details as possible.

The results - in general

• Primary focus is in the center of the image, and only then in the background objects. Large objects, such as buildings, cars, etc, received greater attention.

• There is a tendency to identify people and animals first, unless they are really small compared to other objects in the picture. During the experiment we noticed that sometimes the participants were confused between the type of the animal or the human kind, but in 90% of the pictures that people or animals appeared, there were identified by the participants.

• Once participants identify people or animals, they were primarily focusing on their activities, and only then the rest of the details in the image (about 50% managed to focus on details other than to people or animals). For example: people playing soccer, people smiling, dog holding a Frisbee in his mouth, etc.

• When the image was composed from many small details, participants tended to identify the image scene. For example: beach, room, kitchen, etc.

• There is a tendency to identify things that are brighter. For example: a white lamp. For more information about this result look at the article: “What do we perceive in a glance of real-world scene” by Li Fei-Fei, Asha Iyer, Christof Koch, Pietro Perona.

• 100% (!!!) from the participants identified the following famous places: Eiffel Tower, Golden Gate Bridge in San Francisco and Sydney Opera House. 85% identified the Pyramids. When identified a famous place, the participants didn’t gave attention for details in the picture, such as vegetation, small buildings on the horizon, fountains, sea, etc. In the picture of the Pyramids, only 50% from the participants identified the men and camels, compared to the second section.

The results - specifically by image

In this image 75% of subjects did not identify the flowers in the center of the image. This strengthens the section talks about identifying people and animals. Most of the participants have focused on the people, what they were doing and where they looked. We noticed that participants that did identified the flowers, identified the different vegetation through the all experiment, so we attribute this recognition to the connection that these participants probably have to trees, flowers, grass, etc.

Although this is not a famous site, 90% of participants identified the railroad tracks. Maybe we can connect this to the fact that this is a picture that can look like an image from the Holocaust. Also, many participants connected between the men (father and children, grandfather and grandchildren).

Surprisingly, 75% of the participants did not identify the diver. We can connect this to the fact that this is marine environment, and the participants did not expect to see people at sea. 65% of the participants identified the reefs, which supports the section talks about identifying large things in the front of the picture.

Number of participants imagines things. 10% imagine a cobra in the right image. 10% imagine a computer in the central image. 5% imagine people in the below picture. We tend to link these things to identify the scene by the participants, and connection between things that they would think it makes sense to appear in this scene.

Summary

During this experiment we learned about the detection of the participants, mainly to identify the existence of people and animals in the various images. In addition, we were surprised by the perfect detection ability of the famous sites, even when the time allotted was 35 milliseconds. We hope that future similar trials will take place, perhaps on larger groups of participants. We recommend a similar experiment with color pictures, which number of participants said that can help identify objects. We want to thank Ilan Kedar for the great help in preparing the project.

More information