Note to Selfie: This Algorithm Will Improve Your Picture-Taking Skills
The number of photographs we’re taking is skyrocketing, thanks to new smartphones with powerful cameras and easy access to photo sharing sites. While the quantity of pics has increased, the quality has not kept pace. Resisting the urge to take a “selfie” is a sure-fire way to boost your amateur photog cred, but another path awaits by way of advanced algorithms.
People are using their smartphone cameras to take and share billions of pictures every day. Whether it’s a kid, a cat, a meal or themselves, the pics absolutely must go up on Instagram, Flickr, WhatsApp, Twitter, or Facebook so friends and followers can partake of the joy.
The problem is, most of these pictures are mediocre at best. Sure, the resolution is great with 18-megapixel sensors on the latest Android phones. But when it comes to taking pictures that are pleasing to the eye, the skill of the camera operator nearly always trumps the capability of the camera.
This is where Penn State professor James Wang enters the picture. Wang led the development of a novel approach to improving our photographic skills, and therefore the aesthetic quality of the pictures that we absolutely, positively, must share with everybody via the Internet.
“People are taking a lot of pictures on mobile phones,” says Wang, who is a professor in Penn State’s College of Information Sciences and Technology. “We hope they can leverage technological advancements to improve their picture-taking.”
Wang’s approach involves a mix of automated image analysis by a computer and the aesthetic eye of humans. It starts at Photo.net, a website where amateur and professional photographers share, comment, and rate each other’s photos.
Wang and his colleagues analyzed the picture rankings at Photo.net, and found a strong correlation between the aesthetics and originality ratings for a given photo. “A very original concept leads to good aesthetic value, while beauty can often be characterized by originality in view angle, color, lighting, or composition,” they wrote in a paper describing their work.
This analysis about what is pleasing to the human eye formed the basis for a computation model developed by Wang and his associates. At the heart of the model is an algorithm that will rate the aesthetic qualify of the picture against the standards gathered from Photo.net.
This algorithm formed the basis of a computer program that rate the aesthetic quality of a picture. The software compares the attributes of the new picture against the rated photos, and generates a rating based on a number of attributes, including color, saturation, depth of field and shape convexity.
This rating is provided as instant feedback to the photographer, who can then “learn” from the program about what makes a good picture, according to the collective wisdom of the group of amateur and professional photographers. Based on the feedback, photographers can change their angles or camera settings to achieve a better composition.
Wang and his colleagues were recently granted two patents by the U.S. Patent and Trademark Office (USPTO), and are interested in licensing the technology, possibly to smartphone manufacturers or companies that develop search engines. It could also be used by photo-sharing websites to aesthetically rank pictures.
“The digital photography revolution is likely just at its early stage,” Wang said. “Through harnessing the vast user-generated data on the Internet, future cameras can be equipped with some professional-level knowledge about photography and be able to provide intuitive on-site guidance to photographers.”
Also, please stop with the selfies.