Lesson Content

Image-to-Image generation basics.

Image-to-Image (Img2Img) in Leonardo AI uses an uploaded image as a visual reference that influences the new generation – the model generates content that combines the prompt direction with the visual information from the reference image. The degree of influence is controlled by an Image Strength (or Denoising Strength) slider.

The Image Strength slider.

Image Strength controls how much the reference image influences the output relative to the prompt. Low strength (closer to 0): the output closely follows the reference image – the prompt makes minimal changes to the reference composition and visual character. High strength (closer to 1): the prompt has more influence and the reference image has less – the output changes significantly from the reference toward the prompt direction. For light stylistic changes on a reference: lower strength. For significant creative departure from the reference: higher strength.

What Image-to-Image guides effectively.

Img2Img is effective for: transferring compositional structure from a reference to a new generated image, applying a style transformation to an existing image (changing art style while keeping the composition), creating variations on a specific visual with controlled deviation, and using a rough sketch or concept as a compositional starting point for a more polished generation.

Selecting reference images.

Effective reference images: clearly represent the compositional structure or style being referenced, high resolution, and unambiguous visual content. For composition reference: the composition should be the same as intended. For style reference: the style should be distinctive and consistent throughout the image.