RemapStretch

Documentation

• Class name: RemapStretch
• Category: Bmad/CV/Transform
• Output node: False

The RemapStretch node applies a stretching transformation to images, adjusting their geometry based on specified power factors along the x and y axes, centered around a given point. This node is designed to manipulate image dimensions in a non-linear fashion, offering a way to simulate or correct distortions.

Input types

Required

• power_x
  • Specifies the power factor for stretching along the x-axis. A higher value results in more pronounced stretching.
  • Comfy dtype: FLOAT
  • Python dtype: float
• power_y
  • Specifies the power factor for stretching along the y-axis. Similar to power_x, it controls the degree of stretching but in the vertical direction.
  • Comfy dtype: FLOAT
  • Python dtype: float
• center_x
  • Determines the horizontal center point of the stretch effect. Values range from 0 to 1, representing a percentage of the image's width.
  • Comfy dtype: FLOAT
  • Python dtype: float
• center_y
  • Determines the vertical center point of the stretch effect. Values range from 0 to 1, representing a percentage of the image's height.
  • Comfy dtype: FLOAT
  • Python dtype: float

Output types

• remap
  • Comfy dtype: REMAP
  • The output is a set of parameters or a transformation map that defines how the image should be stretched according to the input parameters. This map is used to adjust the image geometry for stretching effects.
  • Python dtype: Dict[str, Any]

Usage tips

• Infra type: CPU
• Common nodes: unknown

Source code

class RemapStretch(RemapBase):
    @classmethod
    def INPUT_TYPES(s):
        return RemapPinch.INPUT_TYPES_DICT

    def send_remap(self, power_x, power_y, center_x, center_y):
        from .utils.remaps import remap_pinch_or_stretch
        return ({
                    "func": remap_pinch_or_stretch,
                    "xargs": [(1 / power_x, 1 / power_y), (center_x, center_y)]
                },)