Google AI just took photography to a whole new level. The company’s latest AI experiment is taking in images from a Street View car and transforming them into pro-quality landscape images through post processing.
On Thursday, Google’s research lab published Creatism, the results of an experimental deep learning program to test the ability of an artificially intelligent computer to create subjective art.
“Machine learning excels in many areas with well-defined goals,” researchers Hui Fang and Meng Zhang wrote. “However, a clear goal is usually not available in art forms, such as photography. The success of a photograph is measured by its aesthetic value, a very subjective concept.”
To conduct the experiment, Google AI took a large set of images from Google Street View cars. First, the computer had to crop a particular view from the wide range of data and then compose it. According to Google, its machine learning system wandered through about 40,000 panoramas in various locations for landscape photography, which includes California’s Big Sur, Yellowstone, Alps, Canada’s Banff and Jasper national parks.
With a goal of creating a professional or semi-professional image, the program was then asked to edit those images, cropped from 360 panoramas. The program edited those images using digital filters and content-aware lighting adjustments.
Now it’s time to rate what Google AI has created. When presented the final results to actual human photographers, 40 percent of the submissions were rated as professional or semi-professional photography.
The most interesting part of the artificially intelligent system is its ability to apply contextually meaning adjustments in several parts of each and every photograph and create more compelling images. This is, no doubt an incredible job done by Google. The only concern is what should photographers do now?