TEHRAN (ISNA)- An Iranian researcher from Toronto University has designed a new algorithm which has the power to improve the process of searching through billions of photos on the Internet by establishing your relationship with others based on the location of “tags” in photos posted to social media sites.
The tool, created by University of Toronto’s Prof. Parham Aarabi and his former student Ron Appel, uses the proximity of tag locations in photographs to understand the relationships between individuals.
“Essentially, we found that if people are standing close together or are tagged close together inside images – in one image it doesn’t tell you a lot of information. But across hundreds of images that someone has on Facebook, it’s a very good indicator of how close they are in real life, in a social sense,” Aarabi told CTVNews.ca in a phone interview.
The algorithm works by looking at what photos you’re tagged in and what other objects and people are tagged in the photo as well, he said. Based on the results, researchers can produce graphs of how much affinity each person has to other people based on how close they are in the photos.
The algorithm can help accurately match a person’s family and social circle based off of photos posted to social networks, he said.
It even works in photos where an individual is not tagged.
“Imagine you and your mother are pictured together, building a sandcastle at the beach. You’re both tagged in the photo quite close together. In the next photo, you and your father are eating watermelon. You’re both tagged. Because of your close ‘tagging’ relationship with both your mother in the first picture and your father in the second, the algorithm can determine that a relationship exists between those two and quantify how strong it may be,” he said in a statement.
“In a third photo, you fly a kite with both parents, but only your mother is tagged. Given the strength of your ‘tagging’ relationship with your parents, when you search for photos of your father the algorithm can return the untagged photo because of the very high likelihood he’s pictured.”
Aarabi said the tool has the potential to improve the process of searching through photos posted to popular photo sharing sites like Facebook, Flickr and Instagram, because it will turn up photos you weren’t tagged in, but in which you may appear in or that may be relevant to you.
Aarabi notes that there are valid privacy concerns related to this algorithm, but like any new technology, it depends on how it’s used.
The tool will allow you to search for photos of yourself in which you weren’t tagged and you may not have known existed, which is certainly your right, he said. But it will also allow others to search for photos of you that you may not have wanted to be identified in.
“That certainly is a privacy issue and it’s very valid. This, maybe, should not be applied in those cases,” Aarabi said. “It needs to be taken and used with care certainly, but there are cases where it can be useful.”
Another potential use for the algorithm could be map generation – something that was discovered by Aarabi and his team when they took photos of the U of T campus and tagged a few buildings in each photo.
When the photos were put into the system, a graph was produced showing the relationships between the tagged buildings that closely resembled the actual physical map of the campus, Aarabi said.
Eventually, Aarabi wants to make the algorithm – which only uses tag information — smarter, by combining it with visual information such as color, pixel and facial recognition technology.
The Iran Project is not responsible for the content of quoted articles