The Art and Style

In this article I am going to talk about the style and art of images that can be created in AI.

Style

Style is the simplified or stylized form of a represented image of the world. Within the AI these styles are represented by aesthetics tags such as cartoon, realistic, 3d, and anime. In Danbooru you can consult in his tag group the different styles, including abstract, Gothic among others.

You will also find the styles of artists, series and movies. In the tag group we can consult in the list of artists the names of some of them to emulate their style, (it should be mentioned, that the generated image does not belong to the artist himself, but is based on his style). As for series and movies, it can be consulted on the Danbooru page, simply by typing the name of a character, and consult its tag group of the selected image, for example: Judy Hopps, when searching for it we will see that at the top of the list we find Zootopia and also Disney, when using them we will ask to generate a Judy with her original design, however, if we use zootopia and disney, for a character different from Judy we will be able to generate stylized characters as is the zootopia design.

The same happens if we want to generate girls in the style of Pete Peg, it is enough just to place certain characteristics of the character, such as the nose with a specific color such as black nose, or pink nose, and at the end we put Disney or the name of the series "goof trop", and when generating we will get a character based on Peg Pete. In this way we ask to generate a great diversity of styles, it is simply to place the one that pleases us the most and we will be able to generate our characters with more stylization and unique.

Art

Art is the proper representation of the world around us. Many confuse art with the style and the way in which it is made, giving to say that art only exists if it was re-evaluated by a person and not a machine like AI, however, art can be represented in many ways, since the rules are the same.

Light, shadow, colors, perspective, and quality are the fundamental parts of art and knowing how to use them we can generate a true work of art.

Masterpiece, hires, and high quality were initially used in the sd1.5 and sdxl models, due to their instability when generating images and until today many recommend them in pony and illustrious models, without knowing that these indications in these models, are no longer necessary.

Knowing that the Pony and Illustrious models were trained with the Danbooru tags, and when consulting in their tag group, we find a section called composition, in this we find the perspective tags, lighting, camera views, approaches, and color styles, when using any of these tags we will ask to generate an image that is shocking and artistic enough that is desired. As for the simplent quality it is not valid, since these middles generate images with mut good quality.

Conclusion

You can generate images with a style and art with very good quality, since we have the necessary guides to re-evaluate it, unlike other models that only use a natural language, in the pony and illustrious, through the Danbooru tags, we can generate real works of art.