Scientists at MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) have developed a predictive algorithm that will help determine whether a particular photo is memorable or not. The MemNet brain-like computer system uses artificial intelligence to identify parts of a photo that leave the most impression and predicts whether a person will remember a photo or forget it completely.

According to Elle, the algorithm has “near-human levels” of judgment. It has implications beyond simply making one’s photo stand out from others’. The algorithm can be used to develop effective advertising and marketing strategies, teaching resources and memory improvement techniques. With more focus on research, the algorithm might have the power to capture most important information and also store information that human mind is most likely to forget.

Instagrammers and selfie-lovers can use the LaMem demo tool by CSAIL just to have a feel of the algorithm and check whether a photo is going to be memorable or not when posted on social media sites. . LaMem is the world’s largest image-memorability dataset.

The moment one uploads a photo, the tool generates a heat map showing the most memorable areas of the photograph. It also gives the user a memorability score. Apparently, even beautiful landscape photos lose to subject-focused images. Thus, broadly selfies are more memorable than a breathtaking view, as per the tool.

The Sydney Morning Herald reports that the database has over 60,000 images and each and every image has detailed metadata and is carefully annotated. Warm colours on the generated heat map are the memorable parts whereas the cooler colours denote areas not so memorable. From this technology, researchers hope to create an app that can tweak photos to make them more memorable. This can potentially bring in more Instagram likes and may also attract more suitors on dating sites.

The MemNet system uses techniques from a field of artificial intelligence called “deep learning” that uses neural networks to teach computers to process massive data and form pattern on their own. Such techniques are also found in Google auto-complete, Apple Siri and Facebook photo-tagging. They have also made corporate giants invest millions of dollars in “deep learning” start-ups.

Contact the writer of this story at feedback@ibtimes.com.au or let us know what you think below.

RELATED STORIES