Appreciation of AI and Art | Hackaday
In 2019, using Finally, AI for evaluating works of art is more productive than stupid. We all hope that one day soon our Roomba will judge our lifestyle and give unsolicited advice on how we could spruce things up with some pictures and some natural light. There is already a large amount of deep learning dedicated to photo recognition, but a team in Croatia is adapting it for use in fine art. It makes sense that everything is geared towards cameras, as most of us have a huge photo portfolio, but fine art takes longer to render. Even so, the collection on Wikiart.org is extensive and already a hotbed for computer classification work, so they set to work there.
As they modify existing convolutional neural networks, they verify themselves by comparing the results with human evaluations to keep what works and reject what fails. Fortunately, the fine art has a lot of study and commentary extant, while the majority of public domain photographs have nothing more than a filename and maybe some EXIF data. The difference here is that photo analysis AI can say, “This is a STOP sign,” while fine art AI can say, “This is a memorable painting of a sign. “ Everyone agrees that there are still a lot of unknowns when it comes to art and how humans enjoy it, but maybe we are wrong and the reproducible results from computers are correct.
For quick reference, this study found that aesthetics are highest in photos with content and lighting. Bright colors and harmony evoke sentimental feelings. The most memorable paintings focus on the subject. Hope this helps you select your next blog image banner, but we’ve gone for a functional image because that’s what we are.
Here are some of those photo recognition software that prey on Mark Zuckerberg and a hacker who makes himself invisible.
Going through IEEE spectrum.