AILab_ICLoRAConcat Node Documentation

Overview

The AILab_ICLoRAConcat node is designed to facilitate the process of concatenating two images in a specified layout. This node is particularly useful for combining a base image with an object image, using masks to define the blending regions for each image. It allows users to specify the layout direction and customize the size of the output image, providing flexibility in image processing workflows.

Functionality

• Concatenation: This node concatenates an object image with a base image in either a "top-bottom" or "left-right" layout.
• Masking: It utilizes masks to define which regions of each image should be blended in the final output.
• Scaling: The node provides options to resize the images by setting the base image dimensions, ensuring the object image scales proportionally in the specified concatenation direction.

Inputs

1. Object Image (object_image):

   • This is the primary image that forms the foreground (object) in the concatenation. If the image includes an alpha channel, it will automatically be used as the object mask unless a separate mask is provided.

2. Layout (layout):

   • Specifies the direction in which the images are concatenated. Options include "top-bottom" and "left-right."

3. Custom Size (custom_size):

   • An integer specifying the base image size in the resultant concatenation. If set to 0, the node maintains the original size. When a value greater than 0 is specified, it sets either the height (for "left-right") or the width (for "top-bottom") of the base image.

4. Object Mask (object_mask): (Optional)

   • Defines the region of the object image that should be blended into the base image. If not provided, and if the object image is in RGBA format, the alpha channel will be used.

5. Base Image (base_image): (Optional)

   • The background image for the concatenation. If provided in RGBA format and no mask is supplied, its alpha channel will serve as the base mask.

6. Base Mask (base_mask): (Optional)

   • Specifies which areas of the base image should merge with the object image. If not provided, a default mask covering the whole area is used.

Outputs

1. Image (IMAGE):

   • The final concatenated image as per the specified layout and masking regions.

2. Object Mask (OBJECT_MASK):

   • The mask used for the object image indicating which parts were included in the concatenation.

3. Base Mask (BASE_MASK):

   • The mask used for the base image in the concatenation, showing the included regions.

4. Dimensions (WIDTH, HEIGHT, X, Y):

   • Provides the width and height of the output image, along with the x and y positions produced during concatenation.

Use Cases in ComfyUI Workflows

• Combining Portraits with Backgrounds: Users can easily combine an image of a person with another background image, while controlling which portions of each image blend together using masks.

• Creating Extended Canvases: Extend images by adding additional elements to the side or bottom, useful in graphic design tasks like creating banners.

• Enhanced Layer Control: Seamlessly control layer visibility and blending in compositing tasks by defining clear object and base regions using masks.

Special Features and Considerations

• Automatic Mask Extraction: When the inputs include images with an alpha channel and no separate mask is provided, this feature is automatically used as the respective mask.

• Customizable Output Size: The node is flexible with size adjustments, allowing users to maintain resolution integrity or adjust the concatenated output to specific dimensions.

• Versatile Layout Options: The capability to choose between different directional concatenations provides users with versatile design options.

• Minimal Inputs for Simple Tasks: While it offers extensive options, the node can function efficiently with minimal input for straightforward concatenation tasks.

The AILab_ICLoRAConcat node is a powerful tool within the ComfyUI system for users looking to effortlessly blend and concatenate images with precision control over layout and scaling.