The software that fixes your dodgy holiday snaps

Be it a stray thumb or a passer-by blundering into shot, few holiday snaps turn out picture perfect. But help is at hand for even the most amateur of photographers in the form of two computer programs which claim to fix any ruined image within seconds.

The first program - dubbed the Scene Completion software - allows photo-graphers to remove unwanted people or objects from their digital snaps and replace them with perfectly matched alternatives.

Taking into account key measurements, such as camera angles and lighting conditions, the program looks through more than two million photographs on Flickr, an image- sharing website.

It then selects a range of pictures that can be blended on to the original photograph.

So a stray thumb ruining a beach shot could make way for a bucket and spade, or a building blocking a coastline view be removed in favour of miles of glistening sea.

The second approach uses Internet-based image libraries to create a clip-art of objects that, once inserted into a photograph, look convincing.

The software estimates the height of the camera used to take the picture by working out the distance between objects in the snap.

The light sources in a scene are then worked out by looking at the distribution of colour shades within three broad regions; ground, vertical planes and sky.

With knowledge about the position, pitch and height of the camera and light sources, it looks for images in the clip art database taken from similar positions.

The result is so good that during a series of trials the altered snaps passed as originals 70 per cent of the time.

Link & Image: DailyMail
Tags:

Comments

Anonymous said…
But wasn't Photoshop created for this sort of thing in the first place.

Seems to me like re-inventing the wheel
Spluch said…
Yes, but manually using Photoshop to achieve the same effect is a often tedious process. Here's a quote from the original source:

The researchers say that although editing programs such as Photoshop already allow snaps to be tweaked and elements erased or replaced they are often time-consuming and the results unconvincing.

Jean-Francois Lalonde, who headed the research, said: "Matching an object with the original photo and placing that object within the 3D landscape of the photo is a highly complex problem.

"But with our approach, we can hide the complexity from the user and make the process simple and intuitive."