Phys.org: Colorizing images with deep neural networks . “For decades, image colorization has enjoyed an enduring interest from the public. Though not without its share of detractors, there is something powerful about this simple act of adding color to black and white imagery, whether it be a way of bridging memories between the generations, or expressing artistic creativity. However, the process of manually adding color can be very time consuming and require expertise, with typical professional processes taking hours or days per image to perfect. A team of researchers has proposed a new technique to leverage deep networks and AI, which allows novices, even those with limited artistic ability, to quickly produce reasonable results.”
You can have trouble distinguishing colors on the Web but not be colorblind. A new tool aims to held designers identify which color palettes are most accessible. “Christopher Brooks, [University of Michigan] research assistant professor of information, and colleagues surveyed more than 30,000 people to examine how dim conditions and bright sunlight, in addition to varied abilities, can influence how people differentiate color. The researchers then used the survey results to develop ColorCheck, a web tool that can help digital designers see what colors large swaths of their audiences can’t. The software compares color pairs and tells designers what portion of an image’s hues certain percentages of the population can and can’t tell apart. ColorCheck also pinpoints trouble spots on an image by laying a mask of black pixels over them.”
A researcher is working on a way to have a neural network colorize black and white photos. “In a new paper titled ‘Colorful Image Colorization,’ UC Berkeley computer vision PhD student Richard Zhang and his team share how they’re using a ‘convolutional neural network’ to create automatic colorizations that can often fool humans.” It’s not perfect — read the article to see how perfect it isn’t — but for a beginning technology and for automatic coloring, it’s remarkable.
The British Library’s Catalogue of Illuminated Manuscripts has added several new entries. Even if you’re not particularly interested in manuscripts in general, please visit the blog post. These manuscripts are stunning.
More Google: YouTube will soon support HDR video. “Similar to HDR mode on smartphone cameras, which combine multiple exposures into a single frame, HDR video brings out better color and detail. Whereas contrast focuses on bright whites and black blacks, HDR is about all the colors and shades in between, making them more vivid and subtleties in the image more discernible.”
Hey! search for artwork by color. Over two hundred different colors! “To develop the technology, [Alexandra] Chemla says one of her engineers, who has a PhD in mathematics, spent months studying color theory to create a color-detecting algorithm that sorts through hundreds of pixels within each image.” This search tool is a larger part of a site called ArtBinder so the artwork here may not be what you expect, but the color search works far better than I thought it would.