Post Description

Photo collages celebrate important events and themes in our lives. Pick a folder, press a button, 
and in a few minutes AutoCollage presents you with a unique memento to print or email to your Registration
family and friends.
In just a few clicks, people can transform an unwieldy group of photos, for example, of an event 
like a wedding or holiday, into a single collage image that can be printed, used in e-postcards, 
web pages, desktop backgrounds, or emailed to family and friends.
Available to trial from the Microsoft Research download site, AutoCollage is an advanced computer 
vision and image processing program that assembles pictures into a seamless collage using methods 
such as object recognition and face detection. Having automatically identified interesting parts 
of pictures, AutoCollage combines the separate pictures into a composite collage by deliberately 
following natural features as a boundary between images.
“The most significant feature that differentiates AutoCollage is that it offers exceptionally 
sophisticated blending technology for photographs, powered by state-of-the-art computer vision 
techniques,"? said Alisson Sol, development manager at Microsoft Research Cambridge. 
“It’s great that we can give everyone the opportunity to play with and use this compelling technology, 
and we’re looking forward to seeing what collages they come up with!”

There are three main criteria that AutoCollage considers to ensure that it chooses the best images 
to feature in the collages it creates:

* Choosing images that are most representative of the overall theme (eg. avoiding duplications, 
choosing people or groups of people)
* Laying out the images in an interesting way (eg. without one image overly dominating the collage, 
ensuring important parts of the photo are visible)

Blending the images together so the joins aren’t visible
“We’ve tested tens of thousands of different collages in the course of our research, 
and it’s really exciting that the positive feedback we’ve received from our user studies shows 
we’ve answered these challenges successfully,” said Carsten Rother, researcher in the Computer Vision 
Group at Microsoft Research Cambridge.