Visit ComfyUI Online for ready-to-use ComfyUI environment
Powerful node for dynamically splitting images into tiles based on dimensions and overlaps, ideal for AI artists and image processing tasks.
DynamicTileSplit is a powerful node designed to dynamically split an image into multiple tiles based on specified dimensions and overlap settings. This node is particularly useful for AI artists who need to process large images in smaller, more manageable sections without losing continuity between tiles. By dynamically calculating the tile coordinates and handling overlaps, DynamicTileSplit ensures that the resulting tiles can be seamlessly merged back together, making it ideal for tasks such as image processing, texture generation, and other creative applications where working with large images is necessary. The node's ability to handle various tile sizes and overlaps provides flexibility and precision, allowing you to tailor the tiling process to your specific needs.
The image parameter represents the input image that you want to split into tiles. This should be a tensor representing the image data.
The tile_height parameter specifies the height of each tile in pixels. It determines how tall each individual tile will be. The default value is 64, with a minimum of 64 and a maximum of 4096. Adjusting this value will affect the number of tiles generated and their respective heights.
The tile_width parameter specifies the width of each tile in pixels. It determines how wide each individual tile will be. The default value is 64, with a minimum of 64 and a maximum of 4096. Adjusting this value will affect the number of tiles generated and their respective widths.
The overlap parameter defines the number of pixels that will overlap between adjacent tiles. This is crucial for ensuring that there are no gaps between tiles when they are merged back together. The default value is 64, with a minimum of 0 and a maximum of 4096. Increasing the overlap can help maintain continuity between tiles but will also increase the number of pixels processed multiple times.
The offset parameter allows you to specify an offset for the tiling process. This can be useful for creating staggered or non-uniform tiling patterns. The offset value is typically an integer that shifts the starting point of the tiling process.
The tiles_tensor output is a tensor containing all the generated image tiles. Each tile is a section of the original image, split according to the specified tile dimensions and overlap settings. This output is essential for further processing or analysis of the individual tiles.
The tile_calc output is a tuple containing the overlap, image height, image width, and offset values used during the tiling process. This information is crucial for accurately merging the tiles back together and ensuring that the original image dimensions and overlaps are preserved.
© Copyright 2024 RunComfy. All Rights Reserved.