IndianTechnoEra: Image Prompting for AI Art | Prompt Engineering: Master the Language of AI

Chapter 10: Image Prompting for AI Art Models

This module explores techniques for crafting effective prompts for AI art generation models like DALL·E, Midjourney, and Stable Diffusion. Learn to create stunning visual outputs through strategic prompt engineering.

1. Style Modifiers

Style modifiers define the aesthetic of generated images by referencing artistic styles, mediums, or visual qualities.

Common Style Categories:

Art Movements

• Impressionism
• Surrealism
• Art Deco

Mediums

• Oil painting
• Watercolor
• Charcoal sketch

Visual Qualities

• Cyberpunk
• Minimalist
• Cinematic

Artists

• Van Gogh
• H.R. Giger
• Studio Ghibli

Basic Prompt:

"A city at night"

Generic city skyline

With Style Modifiers:

"A cyberpunk city at night, neon-lit with towering skyscrapers and flying cars, in the style of Blade Runner 2049, digital art"

Futuristic neon cityscape

Python Implementation (DALL·E API):

import openai

# Basic prompt
basic_response = openai.Image.create(
  prompt="A city at night",
  n=1,
  size="1024x1024"
)

# Styled prompt
styled_response = openai.Image.create(
  prompt=("A cyberpunk city at night, neon-lit with towering skyscrapers "
          "and flying cars, in the style of Blade Runner 2049, digital art"),
  n=1,
  size="1024x1024"
)

# Display results (implementation depends on your environment)
print("Basic image URL:", basic_response['data'][0]['url'])
print("Styled image URL:", styled_response['data'][0]['url'])

2. Quality Boosters

Specific terms can enhance the technical quality and detail of generated images.

Example Progression:

Basic

"A fantasy castle"

Simple castle illustration

Enhanced

"A detailed fantasy castle with intricate stonework"

More detailed castle

Ultra Quality

"A majestic fantasy castle with intricate stonework, 8K UHD, hyper-detailed, cinematic lighting, unreal engine 5 render"

Photorealistic castle

Effective Quality Terms:

8K UHD hyper-detailed cinematic lighting sharp focus unreal engine 5 octane render photorealistic intricate details

3. Weighted Terms

Emphasize important elements by assigning weights to different parts of your prompt.

Standard Prompt:

"A mystical forest with glowing mushrooms and a small cottage"

Balanced forest scene

Equal emphasis on all elements

Weighted Prompt:

"A mystical forest with glowing mushrooms::1.5 and a small cottage::0.8"

Forest with prominent mushrooms

Mushrooms emphasized, cottage de-emphasized

Weighting Syntax Across Models:

Model	Syntax	Example
Stable Diffusion	(word:weight)	(mushrooms:1.5)
Midjourney	::weight	mushrooms::1.5
DALL·E	Natural language	"Focus on the mushrooms"

4. Fixing Deformed Generations

AI art models often struggle with certain elements like hands, faces, and complex perspectives.

Common Issues and Fixes:

Problem: Distorted Hands

"A pianist playing a grand piano, dramatic lighting"

(Often generates extra fingers or unnatural poses)

Solution:

"A pianist playing a grand piano, dramatic lighting, --no deformed hands"

Problem: Asymmetrical Faces

"A beautiful portrait of a woman with freckles"

(Eyes/nose/mouth may be misaligned)

Solution:

"A beautiful symmetrical portrait of a woman with freckles, perfect facial proportions"

Troubleshooting Guide:

Problem Areas

• Hands and fingers
• Facial symmetry
• Text and signs
• Complex machinery
• Perspective in architecture

Fix Strategies

• Use negative prompts (--no [element])
• Specify "perfect", "symmetrical"
• Avoid close-ups of problematic areas
• Use simpler compositions
• Generate at higher resolutions

5. Negative Prompting

Explicitly exclude unwanted elements or qualities from your generated images.

Without Negative Prompt:

"A serene lakeside cabin at sunset"

May include unwanted elements

With Negative Prompt:

"A serene lakeside cabin at sunset --no people, animals, boats, blurry, low quality"

Clean, focused on cabin

Common Negative Terms:

blurry deformed extra limbs text low quality watermark signature cropped

Python Implementation (Stable Diffusion):

from diffusers import StableDiffusionPipeline
import torch

# Load model
pipe = StableDiffusionPipeline.from_pretrained(
  "CompVis/stable-diffusion-v1-4",
  torch_dtype=torch.float16
).to("cuda")

# Generate with negative prompt
prompt = "A serene lakeside cabin at sunset"
negative_prompt = "people, animals, boats, blurry, low quality"
image = pipe(
  prompt,
  negative_prompt=negative_prompt,
  height=512,
  width=512
).images[0]

# Save or display image
image.save("lakeside_cabin.png")

6. Prompt Structure Best Practices

Effective image prompts follow a logical structure to communicate your vision clearly.

Optimal Prompt Structure:

1. Main Subject: Clear description of primary focus
2. Details: Specific attributes and features
3. Composition: Framing, perspective, camera angle
4. Style: Artistic medium, movement, or influence
5. Quality: Technical specifications
6. Lighting: Mood and illumination
7. Color Palette: Dominant or accent colors

Example Breakdown:

"A majestic white wolf standing on a snowy cliff edge (subject),
with detailed fur, piercing blue eyes, and visible breath in the cold air (details),
low angle shot looking up, full body in frame (composition),
in the style of fantasy concept art, digital painting (style),
8K UHD, hyper-detailed, cinematic lighting (quality),
moonlight casting soft blue highlights (lighting),
cool tones with blue and silver accents (color palette)"

Detailed wolf concept art

7. Iterative Prompt Refinement

Rarely does the first prompt produce perfect results. Iterative refinement is key.

Iteration 1

"A futuristic city"

Generic futuristic buildings

Too vague - needs more detail

Iteration 2

"A cyberpunk city at night with neon lights"

Better but lacks depth

Add style and quality

Iteration 3

"A sprawling cyberpunk metropolis at night, neon-lit streets with holographic advertisements, towering skyscrapers with intricate designs, in the style of Blade Runner 2049, cinematic 8K UHD render, hyper-detailed"

Rich, detailed cityscape

Final version

Prompting Checklist

Before Generating Images:

Have you specified the main subject clearly?

Have you included relevant style modifiers?

Have you added quality boosters if needed?

Have you excluded unwanted elements?

Have you weighted important elements?

Have you planned for iterative refinement?