7.7 KiB
7.7 KiB
Documentation
- Class name: ImageBlendAdvance
- Category: 😺dzNodes/LayerUtility
- Output node: False
- Repo Ref: https://github.com/chflame163/ComfyUI_LayerStyle
用于合成图层，允许在背景图片上合成与之不同尺寸的图层图片，并且设置位置和变换。提供多种混合模式供选择，可设置透明度。
节点提供了图层变换方法和抗锯齿选项。有助于提高合成画质。
节点提供了mask输出可用于后续工作流。
Input types
Required
-
background_image
- 背景图像。
- Comfy dtype: IMAGE
- Python dtype: torch.Tensor
-
layer_image
- 用于合成的层图像。
- Comfy dtype: IMAGE
- Python dtype: torch.Tensor
-
invert_mask
- 是否反转遮罩。
- Comfy dtype: BOOLEAN
- Python dtype: bool
-
blend_mode
- 图层混合模式。
- Comfy dtype: ENUM
- Python dtype: str
-
opacity
- 不透明度。
- Comfy dtype: INT
- Python dtype: int
-
x_percent
- 图层在背景图上的水平位置，用百分比表示，最左侧是0，最右侧是100，可以是小于0或者超过100，那表示图层有部分内容在画面之外。
- Comfy dtype: FLOAT
- Python dtype: float
-
y_percent
- 图层在背景图上的垂直位置，用百分比表示，最上侧是0，最下侧是100。例如设置为50表示垂直居中，20是偏上，80则是偏下。
- Comfy dtype: FLOAT
- Python dtype: float
-
mirror
- 镜像翻转。提供2种翻转模式, 水平翻转和垂直翻转。
- Comfy dtype: ENUM
- Python dtype: str
-
scale
- 图层放大倍数，1.0 表示原大。
- Comfy dtype: FLOAT
- Python dtype: float
-
aspect_ratio
- 图层长宽比。1.0 是原始比例，大于此值表示拉长，小于此值表示压扁。
- Comfy dtype: FLOAT
- Python dtype: float
-
rotate
- 图层旋转度数。
- Comfy dtype: FLOAT
- Python dtype: float
-
transform_method
- 用于图层放大和旋转的采样方法，包括lanczos、bicubic、hamming、bilinear、box和nearest。不同的采样方法会影响合成的画质和画面处理时间。
- Comfy dtype: ENUM
- Python dtype: str
-
anti_aliasing
- 抗锯齿，范围从0-16，数值越大，锯齿越不明显。过高的数值将显著降低节点的处理速度。
- Comfy dtype: INT
- Python dtype: int
Optional
- layer_mask
- 层图像的遮罩。
- Comfy dtype: MASK
- Python dtype: torch.Tensor
Output types
- image
- 合成后的图像。
- Comfy dtype: IMAGE
- Python dtype: torch.Tensor
Usage tips
- Infra type: CPU
Source code
class ImageBlendAdvance:
def __init__(self):
pass
@classmethod
def INPUT_TYPES(self):
mirror_mode = ['None', 'horizontal', 'vertical']
method_mode = ['lanczos', 'bicubic', 'hamming', 'bilinear', 'box', 'nearest']
return {
"required": {
"background_image": ("IMAGE", ), #
"layer_image": ("IMAGE",), #
"invert_mask": ("BOOLEAN", {"default": True}), # 反转mask
"blend_mode": (chop_mode,), # 混合模式
"opacity": ("INT", {"default": 100, "min": 0, "max": 100, "step": 1}), # 透明度
"x_percent": ("FLOAT", {"default": 50, "min": -999, "max": 999, "step": 0.01}),
"y_percent": ("FLOAT", {"default": 50, "min": -999, "max": 999, "step": 0.01}),
"mirror": (mirror_mode,), # 镜像翻转
"scale": ("FLOAT", {"default": 1, "min": 0.01, "max": 100, "step": 0.01}),
"aspect_ratio": ("FLOAT", {"default": 1, "min": 0.01, "max": 100, "step": 0.01}),
"rotate": ("FLOAT", {"default": 0, "min": -999999, "max": 999999, "step": 0.01}),
"transform_method": (method_mode,),
"anti_aliasing": ("INT", {"default": 0, "min": 0, "max": 16, "step": 1}),
},
"optional": {
"layer_mask": ("MASK",), #
}
}
RETURN_TYPES = ("IMAGE", "MASK")
RETURN_NAMES = ("image", "mask")
FUNCTION = 'image_blend_advance'
CATEGORY = '😺dzNodes/LayerUtility'
def image_blend_advance(self, background_image, layer_image,
invert_mask, blend_mode, opacity,
x_percent, y_percent,
mirror, scale, aspect_ratio, rotate,
transform_method, anti_aliasing,
layer_mask=None
):
b_images = []
l_images = []
l_masks = []
ret_images = []
ret_masks = []
for b in background_image:
b_images.append(torch.unsqueeze(b, 0))
for l in layer_image:
l_images.append(torch.unsqueeze(l, 0))
m = tensor2pil(l)
if m.mode == 'RGBA':
l_masks.append(m.split()[-1])
else:
l_masks.append(Image.new('L', m.size, 'white'))
if layer_mask is not None:
if layer_mask.dim() == 2:
layer_mask = torch.unsqueeze(layer_mask, 0)
l_masks = []
for m in layer_mask:
if invert_mask:
m = 1 - m
l_masks.append(tensor2pil(torch.unsqueeze(m, 0)).convert('L'))
max_batch = max(len(b_images), len(l_images), len(l_masks))
for i in range(max_batch):
background_image = b_images[i] if i < len(b_images) else b_images[-1]
layer_image = l_images[i] if i < len(l_images) else l_images[-1]
_mask = l_masks[i] if i < len(l_masks) else l_masks[-1]
# preprocess
_canvas = tensor2pil(background_image).convert('RGB')
_layer = tensor2pil(layer_image)
if _mask.size != _layer.size:
_mask = Image.new('L', _layer.size, 'white')
log(f"Warning: {NODE_NAME} mask mismatch, dropped!", message_type='warning')
orig_layer_width = _layer.width
orig_layer_height = _layer.height
_mask = _mask.convert("RGB")
target_layer_width = int(orig_layer_width * scale)
target_layer_height = int(orig_layer_height * scale * aspect_ratio)
# mirror
if mirror == 'horizontal':
_layer = _layer.transpose(Image.FLIP_LEFT_RIGHT)
_mask = _mask.transpose(Image.FLIP_LEFT_RIGHT)
elif mirror == 'vertical':
_layer = _layer.transpose(Image.FLIP_TOP_BOTTOM)
_mask = _mask.transpose(Image.FLIP_TOP_BOTTOM)
# scale
_layer = _layer.resize((target_layer_width, target_layer_height))
_mask = _mask.resize((target_layer_width, target_layer_height))
# rotate
_layer, _mask, _ = image_rotate_extend_with_alpha(_layer, rotate, _mask, transform_method, anti_aliasing)
# 处理位置
x = int(_canvas.width * x_percent / 100 - _layer.width / 2)
y = int(_canvas.height * y_percent / 100 - _layer.height / 2)
# composit layer
_comp = copy.copy(_canvas)
_compmask = Image.new("RGB", _comp.size, color='black')
_comp.paste(_layer, (x, y))
_compmask.paste(_mask, (x, y))
_compmask = _compmask.convert('L')
_comp = chop_image(_canvas, _comp, blend_mode, opacity)
# composition background
_canvas.paste(_comp, mask=_compmask)
ret_images.append(pil2tensor(_canvas))
ret_masks.append(image2mask(_compmask))
log(f"{NODE_NAME} Processed {len(ret_images)} image(s).", message_type='finish')
return (torch.cat(ret_images, dim=0), torch.cat(ret_masks, dim=0),)