Thursday, October 8, 2015

Software identifies the photos more ‘clickable’ – The Wall Street Journal Americas

Hundreds of millions of photos and videos flooding the internet daily, a potential gold mine for online businesses attempting to get people to click on the images. But that photos will attract more people and generate the most revenue? As companies that publish images can identify the next viral sensation amid a sea of ​​useless things?

Companies doing business on the internet knows that add images to text extends the engagement. Add a picture to a “tweet,” for example, increases by five times the likelihood of users click on the attached link, according to Shift marketing firm. But the increased volume of photos and videos becomes increasingly difficult selection and find out what more attention call is more a matter of art than science.

Now, a new generation of software can help sites to win a greater share of online advertising and e-commerce sectors, which handle, respectively, US $ 149 billion and $ 1.6 trillion per year. Companies like EyeEm Mobile GmbH and Neon Labs Inc. developed software searching images in search of subtle qualities that attract the eyes of Internet users.

“This technology will transform visual industry,” says Stephen Mayes, who is creative director of the former provider of Getty Images photos and experienced software. Just like Google made the Internet more useful to index the pages online, the new software promises to expand the power of images to index and prioritize photos that draw attention, he says.

These tools are the latest in internet betting companies seeking to raise the number of engagement meters as clicks, comments and time spent on pages of a site. Services like Outbrain Inc. and Taboola Inc. automatically populate pages with content that proved popular online, but they often use “bait” that appeal to the lowest denominator of taste: “You’d be surprised [to know] celebrities who still smoke. “

The new software focuses directly on the demand of visual data. Built with techniques of image recognition to identify objects in photos, EyeEm the program seeks common elements in selected images by professional photographers.

The Neon Labs software, in turn, identifies characteristics of images that have already shown, in neuroscience experiments, be able to stimulate brain activity.

Still, Neon Labs is updating the software to recognize photos of celebrities – or more precisely, anyone whose face appears with extraordinary frequency.

“With the election campaign heating up [in the US], we are working with [Images] of [Donald] Trump and [Hillary] Clinton,” says Sophie Lebrecht, a founder of the company.

Luca Paderni, media analyst at research firm Forrester Research Inc, FORR 00:10 % provides that most online image editors will end up using some form of automation via algorithms in the curation of their collections. Automation tools based on how the mind sees the beauty can provide more interesting visual experiences than those offered by companies selling photo files, he says. But as the flashy headlines of online sites, images can also become commonplace over time.

Founded four years ago in Berlin, EyeEm raised $ 24 million from investors like venture capitalist Peter Thiel. The company, which operates an online imaging market, programmed computers to study 50 000 photos classified by a group of experienced photographers. Called “deep learning”, this approach has found millions of characteristics that distinguish the attractive photos of those less interesting, says Ramzi Rizk, director of EyeEm technology. “People say that beauty is a subjective thing,” says Rizk. “But a lot it can be quantified.”

Two months after the release of the software, users of the online market pictures of “startup” – wanting to compete with archival footage from vendors such as Getty Images – were watching, enjoying and commenting with a 30% higher frequency than before, says Rizk. He declined to comment on whether sales improved with it. The platform has not gathered enough data so you can make a comparison, he says.

The profusion of online videos opens up new opportunities. Videos published by companies or on sites like YouTube often show only a single frame or thumbnail of the video to get the audience to see the whole clip. The software produced by Neon Labs analyzes the videos to find the most attractive picture.

The Neon Labs software is based on principles that Lebrecht learned during the ten years he studied neuroscience at Carnegie Mellon University and the National Science Foundation, both in the United States. The brains of volunteers monitored by images obtained by MRI answered photos in missing information such as a face looking away instead of looking forward, she says. They were also attracted to bright lights and for evidence of emotional connections between people. – Between tens of thousands of other data found in research

“There is an innate need to auto [image],” says .

The Neon Labs reported that its customers and partners, including the Sundance Film Festival and American news outlets she would not identify, increased between 16% and 40% to its revenue from videos due to higher number of clicks. The company declined to disclose their prices.

While technology can encourage Internet publishers to observe the basic instincts, it can also raise the quality of images in interesting ways, says Joshua Benton, director of the Nieman Journalism Lab at Harvard University. “Maybe the algorithms are as subtle as the eye Jony Ive, he says, referring to the legendary industrial designer Apple Inc . <-! -> AAPL -1.99 % ” Maybe it’s better to have a lot of Jony Ives that one “

. <- ->

LikeTweet

No comments:

Post a Comment