Skip to content

Image Style Transfer

Transform photographs into artistic styles using natural language prompts.

Image Prompt Template

Transform this {photograph} into {description of output type}: {Description of objects}. It should have {description of style}

Example

Amazon Nova Canvas Condition Image

Watercolor Mountain Lake

Amazon Nova Canvas Image Prompt Text

Transform this photograph into an artistic watercolor painting: A serene mountain lake at sunset, with tall pine trees reflecting in the still water. It should have soft flowing pigments with visible paper texture throughout, brush strokes on the water, delicate color bleeds and gradients, traditional watercolor technique with natural light and shadow transitions, controlled wet-on-wet effects, preserved highlights, professional art style, masterful brush strokes, vibrant yet realistic color palette, slight pigment granulation, subtle paper showing through thin washes, Andrew Wyeth meets William Turner quality, 8k resolution, artstation trending

Amazon Nova Canvas Image Prompt Negative Text

sharp edges, photorealistic, digital art style

Amazon Nova Canvas Image Generation Settings

controlMode: CANNY_EDGE
controlStrength: .7
numberOfImages: 1,
quality: "premium"
width: 1024
height: 768
cfgScale: 8.0
seed: 42

Amazon Nova Canvas Sample Response

Watercolor Mountain Lake

API Request

import json
import boto3
import base64
from botocore.config import Config


bedrock_runtime = boto3.client(
            service_name="bedrock-runtime",
            region_name="us-east-1",
            config=Config(read_timeout=300)
        )

# Read image from file and encode it as base64 string.
with open("mountain-lake.jpg", "rb") as image_file:
    b64_img = base64.b64encode(image_file.read()).decode('utf8')

response = bedrock_runtime.invoke_model(
    modelId="amazon.nova-canvas-v1:0",
    body=json.dumps({
        "taskType": "TEXT_IMAGE",
        "textToImageParams": {
            "conditionImage": b64_img,
            "controlMode": "CANNY_EDGE", # "SEGMENTATION" or "CANNY_EDGE",
            "controlStrength": .7,
            "text": "Transform this photograph into an artistic watercolor painting: A serene mountain lake at sunset, with tall pine trees reflecting in the still water. It should have soft flowing pigments with visible paper texture throughout, brush strokes on the water, delicate color bleeds and gradients, traditional watercolor technique with natural light and shadow transitions, controlled wet-on-wet effects, preserved highlights, professional art style, masterful brush strokes, vibrant yet realistic color palette, slight pigment granulation, subtle paper showing through thin washes, Andrew Wyeth meets William Turner quality, 8k resolution, artstation trending",
            "negativeText": "sharp edges, photorealistic, digital art style"
        },
        "imageGenerationConfig": {
            "numberOfImages": 1,
            "quality": "premium",
            "width": 1024,
            "height": 768,
            "cfgScale": 8.0,
            "seed": 42
        }
    }),
    accept="application/json",
    contentType="application/json"
)

# Process the response to get the generated image
response_body = json.loads(response.get("body").read())

base64_image = response_body.get("images")[0]
base64_bytes = base64_image.encode('ascii')
image_bytes = base64.b64decode(base64_bytes)

# Write the image bytes to a file
output_file_path = "watercolor_mountain_lake.jpg"
try:
    with open(output_file_path, "wb") as output_file:
        output_file.write(image_bytes)
    print(f"Image successfully saved to {output_file_path}")
except Exception as e:
    print(f"Error saving the image: {str(e)}")

aws bedrock-runtime invoke-model \
  --model-id amazon.nova-canvas-v1:0 \
  --body '{
    "taskType": "TEXT_IMAGE",
    "textToImageParams": {
        "conditionImage": b64_img,
        "controlMode": "CANNY_EDGE", 
        "controlStrength": .7,
        "text": "Transform this photograph into an artistic watercolor painting: A serene mountain lake at sunset, with tall pine trees reflecting in the still water. It should have soft flowing pigments with visible paper texture throughout, brush strokes on the water, delicate color bleeds and gradients, traditional watercolor technique with natural light and shadow transitions, controlled wet-on-wet effects, preserved highlights, professional art style, masterful brush strokes, vibrant yet realistic color palette, slight pigment granulation, subtle paper showing through thin washes, Andrew Wyeth meets William Turner quality, 8k resolution, artstation trending",
        "negativeText": "sharp edges, photorealistic, digital art style"
    },
    "imageGenerationConfig": {
        "numberOfImages": 1,
        "quality": "premium",
        "width": 1024,
        "height": 768,
        "cfgScale": 8.0,
        "seed": 42
    }
  }' \
  --accept "application/json" \
  --content-type "application/json" \
  --region us-west-2

{
  "taskType": "TEXT_IMAGE",
  "textToImageParams": {
    "text": "Transform this photograph into an artistic watercolor painting: A serene mountain lake at sunset, with tall pine trees reflecting in the still water. It should have soft flowing pigments with visible paper texture throughout, brush strokes on the water, delicate color bleeds and gradients, traditional watercolor technique with natural light and shadow transitions, controlled wet-on-wet effects, preserved highlights, professional art style, masterful brush strokes, vibrant yet realistic color palette, slight pigment granulation, subtle paper showing through thin washes, Andrew Wyeth meets William Turner quality, 8k resolution, artstation trending.",
    "negativeText": "sharp edges, photorealistic, digital art style"
  },
  "imageGenerationConfig": {
    "numberOfImages": 1,
    "quality": "premium",
    "width": 1024,
    "height": 768,
    "cfgScale": 8.0,
    "seed": 42
  }
}