AI Image to Image Guide - Transform Photos with AI | Apatero | Apatero Blog - Open Source AI & Programming Tutorials
Generation

Transform Any Image with AI Image-to-Image

Image-to-image (img2img) generation takes an existing image as a starting point and transforms it according to your prompt, preserving the original composition while applying entirely new styles, colors, and aesthetics. Turn a photograph into an oil painting, a sketch into a detailed illustration, or a daytime scene into a moonlit night.

Try Image to Image Now

What is Image to Image?

Img2img works by encoding your source image into a latent representation and then using the diffusion process to partially reconstruct it based on your prompt. A strength parameter (typically 0.3 to 0.9) controls how much the output diverges from the original. At low strength (0.3-0.4), the result closely follows the source's composition and colors with subtle style changes. At high strength (0.7-0.9), the AI takes much more creative freedom, often reimagining the entire scene while keeping the general layout. This makes img2img incredibly versatile, useful for style transfer, concept iteration, reference-based generation, and rapid prototyping of visual ideas.

Available on: All plans (starting at $24.99/mo)
How It Works

Step-by-Step Image to Image Workflow

Follow these steps to get the best results with image to image on Apatero

1

Upload Your Source Image

Select any image as your starting point: a photograph, a rough sketch, a previous AI generation, or any visual reference. The composition and major shapes of this image will guide the output.

2

Write Your Transformation Prompt

Describe the style, mood, or changes you want. For example, 'cyberpunk city at night, neon lights, rain-soaked streets' applied to a daytime cityscape photo. Be specific about the target aesthetic.

3

Set the Strength

Adjust the denoising strength. Low values (0.3-0.5) keep close to the original. High values (0.6-0.9) allow more creative freedom. Start at 0.5 and adjust based on your results.

4

Generate and Refine

Generate multiple variations at different strengths to find the sweet spot between preserving composition and applying the new style. Use your favorite result as input for another round to push the transformation further.

Pro Tips

Tips for Better Image to Image Results

Use these tips to get the most out of image to image on Apatero

1

Start with strength 0.5 and adjust up or down based on how closely you want to follow the source

2

Use a sketch or wireframe as input for maximum creative control over composition

3

Chain multiple img2img passes to gradually transform an image through several styles

4

Combine img2img with ControlNet for even more precise structural control

5

High-contrast source images with clear shapes produce the most predictable transformations

6

Try feeding AI-generated images back through img2img with a new prompt to iterate on concepts quickly

FAQ

Image to Image FAQ

Common questions about image to image on Apatero

Q

What is the denoising strength in img2img?

Denoising strength controls how much the output diverges from the source image. At 0.0, the output is identical to the input. At 1.0, the source is completely ignored. Most useful results fall between 0.3 and 0.8, depending on how much transformation you want.

Q

Can I use a photograph as the source for img2img?

Yes. Photos are one of the best source types for img2img. You can transform a selfie into anime, a landscape photo into an oil painting, or a product shot into a stylized illustration.

Q

How is img2img different from ControlNet?

Img2img uses the source image's pixels directly. ControlNet extracts structural information (edges, pose, depth) and uses that as a guide. ControlNet gives you more precise control over composition while allowing complete freedom in style and color.

Q

Can I use img2img to fix or improve an AI-generated image?

Absolutely. This is a common workflow. Take an AI generation that is close to what you want, feed it back through img2img at low strength (0.3-0.4) with a refined prompt, and the output will be a polished version of the original concept.

Q

What image formats does img2img accept?

Apatero accepts PNG, JPEG, and WebP as input formats. For best results, use PNG to avoid compression artifacts in your source image.

Start Using Image to Image

7-day free trial. No credit card required.

Try Image to Image Free