Jump to content
Sign in to follow this  
yossarianba

Занимљив софтвер

Recommended Posts

Ипак је боље прочитати текст на оригиналној локацији:

http://www.primidi.com/2007/02/07.html#a1748

... јер има гомилу линкова за оне које интересују.

A 'beauty function' for a better look

Computer scientists have already tried to digitally enhance our images. Today, let's look at the 'Beauty Function' developed by researchers at Tel Aviv University (TAU) and who promise a more beautiful you. Their program was based on a survey of 300 men and women who were asked to rank pictures of peoples' faces with varying degrees of beauty. Not only it produces results in minutes, but it works. "Volunteers agreed that 79 percent of time, the effects of the Beauty Function -- which can be applied to both men and women -- made a face more attractive." This software could be first used by plastic surgeons. But this enhancement might soon also become available on your digital camera.

Before going further, below are two photographs of a young woman. On the left is the original input image while the beautified result appears on the right. (Credit: Daniel Cohen-Or and Tommer Leyvand)

Some 'Beauty Function' results

This software has been conceived by TAU's Daniel Cohen-Or and Tommer Leyvand. Here is how they started their project.

In developing the Beauty Function, they surveyed 300 men and women and asked them to rank pictures of peoples' faces with varying degrees of beauty, on an attractiveness scale of 1-7. The scores were correlated to detailed measurements and ratios of facial features such as nose width, chin length and distance from eyes to ears. Some 250 measurement points were taken into account and once formulated, researchers developed an algorithm that could let them apply some of the desired elements of attractiveness -- as mathematical equations -- to a fresh image.

The illustration below describes the full digital face beautification process. (Credit: Daniel Cohen-Or and Tommer Leyvand)

The 'Beauty Function' process

Another quote from the original article from Karin Kloosterman is worth reading.

"Beauty is not in the eye of the beholder," says co-researcher Cohen-Or. "Beauty is merely a function of mathematical distances or ratios. And interestingly," he adds, "It is usually the average distances to features which appears to most people to be the most beautiful. I don't know much about beauty and I don't pretend that I do," he adds, "But the nice thing about this project is that we didn't intend or aim to define beauty. We don't care about the reasons that make someone appear to be more beautiful. For us, every picture is just a collection of numbers."

I advise you to read the full article from Karin Kloosterman: you will not waste your time.

The images on this post have been picked from two documents about "Digital Face Beautification," the first one being a gallery while the other is a short technical paper presented at SIGGRAPH 2006 (PDF format, 1 page, 2.31 MB)

For more information about this software, you can read an older article from New Scientist, "From snapshot to cover model in a single click" (Celeste Biever, New Scientist, August 10, 2006).

Finally, it's not the first time that computer scientists try to digitally enhance our images. For example, here is an article from CNET News.com, "Digital cameras focus on revised reality" (Candace Lombardi, CNET News.com, August 29, 2006), about some digital cameras sold by Hewlett-Packard.

Sources: Karin Kloosterman, for ISRAEL21c, February 4, 2007; and various other websites

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
Sign in to follow this  

×
×
  • Create New...