EasyUse/Imagegenerated

Image Detail Transfer(imageDetailTransfer)

The Image Detail Transfer node transfers fine-grained detail from one image (the source) onto another (the target).

image Detail Transfer

target
source
mask
image
mode
add
blur_sigma
1.00
blend_factor
1.000
image_output
Preview
save_prefix
ComfyUI
Easy Use

Description

The Image Detail Transfer node transfers fine-grained detail from one image (the source) onto another (the target). It is commonly used to bring texture, grain, or high-frequency detail from a detail-rich image into a target image while leaving the target’s overall color and structure intact. The node extracts a detail layer from the source using a Gaussian blur, then composites that detail into the target using one of several blend modes, with an optional mask to limit the effect.

Inputs

NameTypeDefaultDescription
targetIMAGE—The base image that receives the transferred detail. Its colors, composition, and lighting remain the dominant content of the output.
sourceIMAGE—The image from which detail is extracted. This image usually contains the fine texture, edges, or grain you want to layer onto the target. The node blurs the source to compute the detail layer before blending.
modeCOMBO"add"The blend operation used to combine the extracted source detail with the target. Options: add, multiply, screen, overlay, soft_light, hard_light, color_dodge, color_burn, difference, exclusion, divide.
blur_sigmaFLOAT1.0Standard deviation of the Gaussian blur used to separate detail from the source. Small values isolate only the finest high-frequency detail; larger values include broader local contrast and texture. Range: 0.1–100.
blend_factorFLOAT1.0Strength of the transferred detail. 1.0 applies the full extracted detail, 0 leaves the target unchanged, lower values attenuate the effect, and negative values invert the detail contribution. Range: -10–10.
image_outputCOMBO"Preview"Controls display and saving of the result. Options: Hide — no preview or save; Preview — display in the UI; Save — save to disk; Hide/Save — save without displaying.
save_prefixSTRING"ComfyUI"Filename prefix used when saving the output image, for Save and Hide/Save modes.
maskMASK—Optional mask limiting where detail transfer is applied. White areas receive the full effect, black areas keep the target unchanged, and gray values blend partial strength. If omitted, the transfer applies to the whole image.

Outputs

TypeDescription
IMAGEThe resulting image after detail transfer. Its dimensions and overall color layout match the target, with source-derived detail blended in according to the selected mode, blur_sigma, blend_factor, and mask.

Usage Notes

  • The typical workflow is: blur the source to isolate its detail, then composite that detail onto the target using the chosen mode and blend_factor. In add mode this is essentially adding the high-frequency source detail directly to the target.
  • blur_sigma controls the scale of detail being transferred. Use small values (about 0.5–2) for fine texture or sharpening-like detail, and larger values for broader texture patterns or lighting detail. The best value depends on the resolution of your images.
  • If the result is too strong, reduce blend_factor. If the result looks inverted or produces unwanted halos, try a different mode or use a negative blend_factor to flip the detail contribution.
  • The optional mask is useful for applying detail to specific regions, such as skin, fabric, or foreground areas, while keeping other parts of the image untouched.
  • When image_output is set to Save or Hide/Save, the output image is written using save_prefix as the base filename, following the Easy-Use package’s standard save behavior.
  • The node always returns the IMAGE output regardless of the image_output setting, so it can still be connected into the graph for further processing.

Comments

Sign in with GitHub to join the discussion.

Loading comments…