Create series of images with the same style, but different content

API
1

I’d like to create a series of images based on a source image, and this series of images can combine together to tell a coherent narrative. I’d also like each image to maintain the same style (color, theme, etc). The images can combine together to tell a coherent narrative. The images are not a variation of the root, but a totally new image with the same source content.

For example, I’d like to provide a source image of a specific dog, and then generate a series of images that tell a narrative about the same dog, all in the same style.

Dall-e-3 doesn’t have a method to provide a ‘source’. How can I achieve this?

2

DALL-E 3 cannot ingest other imagery. Its only input is words.

Within various iterations of ChatGPT, there’s been the ability for the AI model to specify a seed or a previous image to its version of DALL-E so that future requests can share similar aspects of generation, but the API doesn’t expose such a method.

Instead, you’d have to overwhelm the image generator with stylistic words and then just change small aspects of the contents. Or look at other’s tools.

1 Like