Stable Doodle, a unique sketch-to-image technology that brings basic designs to life, has been launched by Stability AI.
Stable Doodle allows anybody with minimal sketching abilities and internet connectivity to create spectacular pictures in seconds, saving time and boosting productivity. It can be used for everything from designing for customers and producing presentation materials to making website designs and even logos.
Stable Doodle combines Stability AI’s advanced image generation technology, Stable Diffusion XL, with Tencent ARC’s T2I-Adapter. The T2I-Adapter adds control to AI image generation by incorporating trainable parameters into existing models. In Stable Doodle, it improves the pre-trained text-to-image model (SDXL) to understand sketch outlines and generate images based on combined prompts and outlines.
Users may access Stable Doodle via the Clipdrop website and app, both of which are accessible for iOS and Google Play. There is no need to log in, however there are daily use limitations. The user interface consists of drawing a simple sketch, selecting one of Stable Diffusion XL’s 14 art styles, and hitting ‘produce.’
Users may also choose from a variety of genres, including realistic photography, cinematic compositions, imaginative fantasy art, and origami-inspired patterns. However, the output is impacted by the user’s original drawing and description, and the accuracy varies based on the intricacy of the scene.
The sources for this piece include articles in stability.ai and ZDNET.