Most Memorable Photos Are Those of People

In a new study, researchers at the Massachusetts Institute of Technology determined that the most memorable photographs people are make are those containing other people, and not scenic lakes and landscapes. In fact, the latter are utterly forgettable, the team says.

As such, it may be best for people to start photographing each other while on a trip. In the new classification developed at MIT, another top contender for most memorable image is represented by static indoor scenes, followed closely by objects at a human scale.

Magnificent, large-scale landscapes are also beautiful, but they are more a thing of the moment, experts say. They can be very easily forgotten as individuals who go on vacation return to their daily routines.

According to one of the lead authors on the new paper, MIT graduate student Phillip Isola, there is a great difference between pleasantness and memorability. The two definitely don't go hand in hand.

Details of the new investigation will be presented to scientists in Colorado Springs between June 20-25, at the IEEE Conference on Computer Vision and Pattern Recognition. This is the first time such a study is conducted with scientific rigor.

Until now, investigators believed that visual memory was simply too subjective to be analyzed accurately. The MIT team proved that this is not the case, and revealed some interesting correlations in the process.

“People did not think it was possible to find anything consistent,” explains MIT associate professor of cognitive science Aude Oliva, who was also a senior author of the new scientific paper.

The research was in fact so successful that the MIT team, which included Antonio Torralba, the Esther and Harold E. Edgerton associate professor of electrical engineering and computer science, even managed to develop a computer algorithm explaining it.

Future applications for the formula could go a long way towards helping photo editors and graphic designers decide on the content they want to include in their respective products. As far as end-users go, people could soon use new software to tell them which vacation images are most memorable.

In a series of experiments conducted in the lab, participants were asked to rate a number of images, and then the experts looked at the correlations that existed between the results.

“There are always differences between observers, but on average, there is very high consistency,” explains Oliva, who also holds an appointment as the principal investigator in the computer vision group at MIT’s Computer Science and Artificial Intelligence Laboratory.