New Algorithm Predicts the Memorability of Your Images
How memorable are your photos? A computer can tell you just that. Scientists at MIT have developed an algorithm that can predict how memorable or forgettable an image is almost as accurately as humans.
"Understanding memorability can help us make systems to capture the most important information, or, conversely, to store information that humans will most likely forget," said Aditya Khosla, the lead author on a related paper, in a news release. "It's like having an instant focus group that tells you how likely it is that someone will remember a visual message."
The new algorithm is actually set to be developed into an app that subtly tweaks photos to make them more memorable. However, there are many other applications for it.
For example, the algorithm could be used to improve the content of ads and social media posts. It could also be used to help develop more effective teaching resources and to create a personal device to help a person remember things.
The researchers actually pitted the algorithm against human subjects by having the model predict how memorable a group of people would find a new, never-before-seen image. The algorithm actually performed 30 percent better than existing algorithms and was within a few percentage points of the average human performance.
"While deep-learning has propelled much progress in object recognition and scene understanding, predicting human memory has often been viewed as a higher-level cognitive process that computer scientists will never be able to tackle," said Aude Oliva, principal research scientist. "Well, we can, and we did!"
You can actually upload your own images online to see just how memorable they are here. The algorithm will create a "heat map" over the image that identifies exactly how memorable or forgettable an image is.
For more great science stories and general news, please visit our sister site, Headlines and Global News (HNGN).