Sign In

Stable Diffusion : Create a comics page in few seconds

Objective

Create a comics page in a second.

Preface

I am not a comic expert, I am studying techniques with AI, and this is a TEST I did last night.

Environment

For this test I will use:

Checkpoint

I will use a checkpoint for generating nice graphics

  • Aniverse (https://civitai.com/models/107842/aniverse)

Create an example page

After installing Krita.

Run and create a new document with size 512x768

Wih the rectangle tool add some rectangles to generate a comics page.

When you finish save the image in png format (Menu file >> Save )

Attached to this article there is the file MyPage.zip that contains my png .

Generating the comics

Go to stable diffusion

Prompt

comics of a group of friends gathered around a table, Ireland, San patrick day, clinking glasses filled with our refreshing beer. <lora:more_details:1> <lora:people_count_slider_v1:2>

Negative:

blurry, blurry_image, Lowres, Low_resolution, Low_picture_quality, Low_picture_anime, extra_anatomy, extra_body, extra_navel, extra_face, extra_eyes, extra_chest, extra_nipples, extra_hips, extra_arms, extra_hands, extra_fingers, extra_legs, extra_feet, extra_toe, missing_anatomy, missing_body, missing_navel, missing_face, missing_eyes, missing_chest, missing_nipples, missing_hips, missing_arms, missing_hands, missing_fingers, missing_legs, missing_feet, missing_toe, sketch, sketching, (worst quality: 2), (low quality:2), ((monochrome)), ((grayscale)), inpainting, HDR, bad_prompt_version2, jpeg artifacts, signature, watermark, text, error, missing fingers,naked,

I am using couple of loras that I like:

Other

Steps: 25

Sampler: Euler a

CFG scale: 7

Seed: 725189896

Face restoration: GFPGAN

Size: 512x768 (I use Hires, also to increase the size)

ControlNet

I will add the image that I have generated.

The I enable the component and select SoftEdge.

Generate the image

After few try I got some images

These test were made with other prompts:

This one without details.

This is my finale Image with "Add detail " lora and people.

Check my post for the "Celtic Creation content" : https://civitai.com/images/7470797

Regional prompter

Now let's improve the image using regional prompter

Regional prompter permits to give information for each part of the image.

Check out my article about this: https://civitai.com/articles/3959/stable-diffusion-testing-the-regional-prompter-plug-in-using-break

In this case we will use Columns and lines to cover our comics page.

I set the active flag to on, the I select columcs.

I set the size to 512x768

For the divider radio I use 1,1,1 ; 1,1,1 ; 1 => The ; separate the lines, the , (commas) separate the columns.

some example to give you the idea of how it works.

Now let change the prompt accordingly.

Prompt

comics of a group of friends gathered around a table, Ireland, San patrick day, clinking glasses filled with our refreshing beer. <lora:more_details:1> <lora:people_count_slider_v1:2>

BREAK

smiling and drinking beer. Red h<ir

BREAK

A girl in green dress with irish hat looking around

BREAK

The bartender with a huge beer in the pub

BREAK

Zoom of one nice fresh beer

BREAK

1boy and 1 girl looking each other with the beer thinking about their future love

Result

From now on you can work with your prompt:

comics of a group of friends gathered around a table, Ireland, San patrick day, clinking glasses filled with our refreshing beer. <lora:more_details:1> <lora:people_count_slider_v1:2>

BREAK

smiling and drinking beer. Red hair

BREAK

A girl in green dress with irish hat looking around

BREAK

The bartender with a huge beer in the pub

BREAK

Zoom of one nice fresh beer

BREAK

1boy and 1 girl (looking each other:1.1) with the (beer:1.4) thinking about their future love

Check also with other checkpoints to see the results:

Checkpoint: Meinamix

Sometimes the result is not as good as expected, in this case there is no beer in the images

Checkpoint : Real Cartoon 3d

Last test

Checkpoint : Camellia Mix.

For the title of this article I created an image 800x400

and generated the comics

I hope you enjoy this article.

22

Comments