Adobe demonstrates a future Photoshop feature called “Content-Aware Fill” that is the healing brush on crack. It’ll “extrapolate” image patterns to remove unwanted elements or extend landscapes.

Looks pretty amazing in this youtube video:

It’s probably based on some Siggraph paper. For example, the content-aware-resizing that Photoshop has since CS4 was also demoed at Siggraph first:

The algorithms that are currently developed in this field are simply amazing. Here’s the one that will put me out of work in a few years: automatic image composition based on a simple doodle:

Sketch2Photo: Internet Image Montage from Tao Chen on Vimeo.

Or this one on image stabilisation:

And finally this all-time favourite from 2007 (!) that apparently still hasn’t made it into a commercial product: An algorithm that seamlessly inserts high-resolution photos into a low-resolution video. Or removes unwanted elements. Or removes reflections. Or seamlessly changes what was behind those reflections.

Using Photographs to Enhance Videos of a Static Scene from pro on Vimeo.

Tags: ,

Comments are closed.