AI Painting Popularization Course【2】Pictures Make Pictures

3. Pictures and pictures

abstract:

1. Principle of graph-generating graph

2. Basic process of drawing drawings

3. Analysis of the role of random seeds

1. Principle of graph-generating graph

Tushengtu can help you draw a picture into another look. In the Vincentian diagram, we see that the AI ​​Vincentian diagram has a certain degree of randomness, and the things drawn may not fully meet our needs. In real life, this kind of transmission deviation and misinterpretation of requirements is actually common. For example, Tiantian Qianjia’s product managers and programmers, Party A’s customers and designers. When communicating with AI, we can give it a picture as a reference to generate a new picture.

2. Three key steps to create a graph

  • Step 1: Import images
  • Step 2: Write the prompt word
  • Step 3: Parameter adjustment

3. Technical analysis of parameters

Refer directly to the prompt questions on the Vincentian diagram.

Positive prompt words:

(masterpiece:1,2), best quality, masterpiece, highres, original, extremely detailed wallpaper, pertect lighting, (extremel detailed CG:1.2), drawing, paintbrush,

Reverse prompt words:

NSFW, (worst quality:2) , (low quality:2), (normal quality:2) , lowres, normal quality, ((monochrome)), ((grayscale)), skin spots, acnes, skin blemishes, age spot, (ugly:1.331), (duplicate:1.331), (morbid:1.21), (mutilated:1.21), (tranny:1.331), mutated hands, (poorly drawn hands:1.5), blurry, (bad anatomy:1.21), (bad proportions:1.331), extra limbs, (disfigured:1.331), (missing arms:1.331), (extra legs:1.331), (fused fingers:1.61051), (too many fingers:1.61051), (unclear eyes:1.331), lowers, bad hands, missing fingers, (((extra arms and legs))),

parameter settings:

  • Redraw width

    Indicates the proportion of the difference between the drawn picture and the original picture.

  • resolution

    It’s best to have the same proportion as the picture, otherwise it will appear sideways.

4. Research on the meaning of random seeds

The images generated by AI are random, but each generation has its own set of depiction methods, and this depiction method will be recorded as a set of random numbers. We call this a random seed.

Different random seeds will of course produce highly random effects, but if you use the same random seed, there will inevitably be many similarities in the generated images because they are randomly generated using the same method. of.

From its perspective, you are basically right to say that AI painting is a card draw, because every time you generate it, you will get a random seed. Some seeds are good and some are not.

In the random seed setting, there are two function buttons. Click the dice to set the random parameter to -1, which means a new card will be drawn every time. Clicking the cycle button will set the seed to the number you drew on the last card.

Guess you like

Origin blog.csdn.net/xianyu120/article/details/133267057