Create a visual from a reference image
To create an image, you can, of course, use the "text to image" method, which means writing a prompt that describes what you have in mind (see the article
Create an Image with a Text Prompt), but you can also use a reference image to create variations.
This method is very effective for having more control over what you produce, stylizing an image in another style, or creating variations of a generated image.
Note: Even when using a reference image, it is recommended to add a prompt for better results.
To add a reference image, you can either:
* Add an image from the creation flow
Simply click on the "Input Image" button that appears when you hover over the image. The image will then be sent to the "Reference Image" block in the left column, below the model section.
* Upload an image from your computer
Simply click on the image shown above and select the image you want from your computer or mobile device.
The selected image will be automatically resized.
Currently, when you add a reference image, Seelab only uses it to inspire the structure of the added image. Very soon, you will also be able to use an image as a style reference.
When using the structure reference, Seelab allows you to use multiple edge detection algorithms to recreate an image based on the uploaded image.
Each detection type uses specific elements and will be more or less effective depending on the type of image added.
Simple “Canny Edge” : Detects the main lines of an image. The most versatile.
Artistic “Painting” : Detects artistic contours. Ideal for artistic images.
Sketch “Sketch” : Detects the contours of a pencil drawing. Ideal for sketches.
Visual Distance “Depth Map” : Detects whether elements are near or far. Ideal for retaining only the layout of elements in a room, for example.
Human Pose : Detects how humans are positioned. Ideal for reproducing a specific pose.
When using the reference image feature, you can adjust the "balance" slider to define whether the image should be very close to the original image in terms of structure, or if you want Seelab to be more creative and just use your image as a starting point.
In the example above, by pushing the slider towards "closer to the image", the newly generated images will be very close to the original image in terms of structure and pose.
On the other hand, by moving the slider towards "better creativity", Seelab will keep the subject but suggest variations in terms of pose and image composition.
It's up to you to find the right balance according to your needs!
Don't forget - once you're done with your reference image - to remove it so it doesn't affect your subsequent generations!
Create an Image with a Text Prompt), but you can also use a reference image to create variations.
This method is very effective for having more control over what you produce, stylizing an image in another style, or creating variations of a generated image.
Note: Even when using a reference image, it is recommended to add a prompt for better results.
Adding a Reference Image
To add a reference image, you can either:
* Add an image from the creation flow
Simply click on the "Input Image" button that appears when you hover over the image. The image will then be sent to the "Reference Image" block in the left column, below the model section.
* Upload an image from your computer
Simply click on the image shown above and select the image you want from your computer or mobile device.
The selected image will be automatically resized.
Using a Structure Reference Image
Currently, when you add a reference image, Seelab only uses it to inspire the structure of the added image. Very soon, you will also be able to use an image as a style reference.
When using the structure reference, Seelab allows you to use multiple edge detection algorithms to recreate an image based on the uploaded image.
Each detection type uses specific elements and will be more or less effective depending on the type of image added.
Simple “Canny Edge” : Detects the main lines of an image. The most versatile.
Artistic “Painting” : Detects artistic contours. Ideal for artistic images.
Sketch “Sketch” : Detects the contours of a pencil drawing. Ideal for sketches.
Visual Distance “Depth Map” : Detects whether elements are near or far. Ideal for retaining only the layout of elements in a room, for example.
Human Pose : Detects how humans are positioned. Ideal for reproducing a specific pose.
Importance of the Balance between Image Fidelity and Creativity
When using the reference image feature, you can adjust the "balance" slider to define whether the image should be very close to the original image in terms of structure, or if you want Seelab to be more creative and just use your image as a starting point.
In the example above, by pushing the slider towards "closer to the image", the newly generated images will be very close to the original image in terms of structure and pose.
On the other hand, by moving the slider towards "better creativity", Seelab will keep the subject but suggest variations in terms of pose and image composition.
It's up to you to find the right balance according to your needs!
Don't forget - once you're done with your reference image - to remove it so it doesn't affect your subsequent generations!
Updated on: 04/10/2024
Thank you!