Are you thinking of getting a 3D printer for your home or business? Visit Top 3D Shop and read our latest reviews. We will help you get the machine you need!
WordsEye has begun a closed beta test of its service, which allows creating three-dimensional models of the described scenes based on sentences written in natural language. It is reported by TJournal with reference to The Next Web. Samples of the generated work can be found in the gallery on the service website.
In order to create a three-dimensional scene in the WordsEye service, the user does not need to have the skills to work in 3D editors or the experience in modeling three-dimensional objects. The system as input receives from the user a text description of the future scene in free form in English. The algorithm extracts objects from the text, data on their location, coverage, color and lighting, and at the output the user receives a generated three-dimensional scene.
Due to the imperfection of the algorithms for recognizing and processing objects, the system sometimes skips some parameters. In addition, the inability to accurately describe the required scene in all details sometimes leads to unexpected results, leaving the system room for "independence", so most of the finished scenes are similar to work in the direction of surrealism.
Giant gray mouse on a six-meter shiny cube. On the mouse a meter and a half golden lock. There is a yellow light above the castle. A six-meter shiny gray hand a meter above the castle. Gold DNA 16 meters high stands on the ground 60 centimeters from the mouse. Dusk. Grass on the ground.
At the stage of closed beta testing, users replenish the database with all kinds of 3D objects, which the service will subsequently operate on. According to the developers, they position the service not only as an entertainment portal, but also as a means of rapid modeling. In addition to the online service, it is also planned to develop applications for Android and iOS.
