Text-to-image (T2I) translation is a type of artificial intelligence that generates an image based on a written description or a textual prompt. The system uses deep learning algorithms to analyze the text and create an image that represents its content.

The entire process can be divided into the following steps:

• Text analysis: the system analyzes the written text to understand its content, including objects, scenes, and relationships between them.
• Image generation: the system uses pre-trained deep learning models to generate a preliminary image based on the textual description.
• Refinement: the preliminary image is then refined through additional deep learning algorithms to generate a final, high-quality image.

Today’s well-known T2I models like DALL-E 2 and Midjourney use this technology but apply it in different ways to generate unique images. Learn more about how they do it on dedicated Glossary pages.