Software turns grayscale images into realistic colours

Existing methods require the user's input, either by using a scribbling tool to colour the image manually or by using a colour transfer. Both options can result in poor colourisation quality limited by the user's degree of skill or the range of reference images available.

"We have developed a method that takes advantage of the plentiful supply of internet data to colourise gray photos," said Alex Yong-Sang Chia of the Singapore-based A*STAR's Institute for Infocomm Research, who along with colleagues developed the computer program.

The software searches hundreds of thousands of online colour images, cross-referencing their key features and objects in the foreground with those of grayscale pictures, according to an A*STAR statement.

"The user segments the image into separate major foreground objects and adds semantic labels naming these objects in the gray photo. Our program then scans the internet using these inputs for suitable object colour matches," said Chia.

Given the vast amount of visual data available online, not all of the chosen images are useful. Once the initial colour images have been found, the program then filters them to find the most realistic and suitable matches for the grayscale object inputs.

"Our coloured images were classed as 'real' up to 65 percent of the time," says Chia. "Overall the colourisation results are visually pleasing and perceptually meaningful to users."