The Potential of Art Discovery through Artificial Intelligence Innovation

2019-11-18

Artificial intelligence has almost touched every industry and discipline known to mankind, and the field of culture and art is no exception. Artrendex is an AI-driven new interface that, along with its component ArtPI, may potentially change the way art is discovered, exhibited, and sold. ArtPI is the first public API designed and optimized for art. It utilizes artificial intelligence and deep learning models to train on over a million artworks, in order to understand concepts of style, genre, subject matter, composition, lighting, space, color, and other artistic principles and elements. Founded by Ahmed Elgammel, ArtPI leverages centuries of art pieces to detect features and patterns based on overall visual styles derived from art theory and history. By detecting stylistic tags, themes, and artists, it establishes connections within a large collection, providing project planning solutions for the art market, professionals, or businesses. For museums and other institutions, this means that users can visually search their collections and explore artworks in a more intuitive way by searching for specific "tags." Curators can have a comprehensive overview of all the artworks, discovering unexpected connections or similarities between them, thus inspiring new ways of presenting or interpreting the works. For galleries, dealers, and auction houses, ArtPI can be combined with social media to reveal trends or predict market conditions. As it is specifically built for art and centuries of art pieces, using an overall visual style based on art theory and history to detect feature patterns, it can visually find artworks with similar stylistic tags and creation periods, identify themes and artists, and establish connections within a vast collection. Even if the images of artworks are distorted, ArtPI can still work. It allows for less-than-ideal source images to be captured from various angles, as the AI model can be used to correct distortions and extract information that can be compared with other artworks. This eliminates the need for complex photography requirements and enables ArtPI to work in various imperfect environments. When it comes to recognizing visual patterns, artificial intelligence often outperforms humans. ArtPI utilizes neural networks to train its algorithms on images, starting with over 250,000 artworks. The more tasks the algorithm encounters, the more subtle the results become. Elgammal states, "Humans can quickly understand the meaning of text, which is something that artificial intelligence cannot match. However, it can find millions of examples in a matter of seconds or fractions of a second. Each search and each image added to ArtPI will contribute to its improvement, and over time, its value will continue to increase." Note: The HTML tags and structure have been preserved as requested.

RECENT AI NEWS

RECENT AI TOOLS