Controlnet Image To Art Model Guide

Prompt used to generate the above sample:

"RAW photo, portrait photo of 30 y.o woman queen, pale skin, slim body, (high detailed skin:1.2), background ocean, 8k uhd, dslr, soft lighting, high quality, film grain, Fujifilm XT3"

Introducing controlnet models that combined with stable diffusion and other intermediate AI models to support generating art from a photo or image

Creata AI Controlnet+Stable Diffusion Models:
- Canny: combination of stable diffusion + controlnet
- Deliberate: it's great at generating an art, avatar or photo based on a portrait of a person. Compare it with Lensa.
- Struture: great at generating interior design images by detecting strutural information from the image
- HED: Controlnet with HED image detection
- Realistic: Generate realistic art images

Models

Controlnet Canny

Use canny model for image edge and feature detection, then employ stable diffusion for final art generation.

Controlnet + Deliberate

Use controlnet canny for image detection then employ Deliberate model based on stable diffusion for final art generation.

Controlnet Structure

The model makes use of hough image detection algorithm which is good at detecting structural elements in the input image. It's great in generating interior design or architectural images.

Controlnet HED

HED creates image that is faithful to the input image or photo. One can use it to tur your photo into a fantacy art.

Controlnet Realistic

This model combines the power of controlnet and realistic vision to create amazing art effects.

Parameters

Some of the models provide input Parameters for fine tunning image generation. Here are some of the most common model parameters you should know about:

• Scale/Guidance: a value between 0.1 - 30. Used to specify how closely the generation should be to the prompt's description. The larger the value closer the result image to the prompt
• weight/stength: a value between 0.01 - 2, the weight that controlnet model influences art generation
• distance threshould: a value in the image the model should use to detect input image features. If you want it to pick image elements farther away, use a larger value.
• Color value threshould: the image colors that are larger than the specified value will be detected. The larger the value is, the more brighter part of the input image will be used in the final art image creation.