111 lines
3.3 KiB
Markdown
111 lines
3.3 KiB
Markdown
|
|
# Documentation
|
|
- Class name: EnhanceDetail
|
|
- Category: image/filters
|
|
- Output node: False
|
|
|
|
EnhanceDetail节点旨在通过应用一系列滤镜来提高图像的视觉质量，这些滤镜可以增强细节、调整锐度，并可能降低噪声。该节点通过相对于引导滤波器输出来操作图像的细节，从而对增强过程实现精细控制。
|
|
|
|
# Input types
|
|
## Required
|
|
- images
|
|
- 需要增强的输入图像。这个参数至关重要，因为它直接影响增强过程，作为所有后续操作的基础。
|
|
- Comfy dtype: IMAGE
|
|
- Python dtype: torch.Tensor
|
|
- filter_radius
|
|
- 指定增强过程中应用的滤镜半径。较大的半径可能会导致更明显的细节增强，但也可能增加处理时间。
|
|
- Comfy dtype: INT
|
|
- Python dtype: int
|
|
- sigma
|
|
- 控制引导滤波器的sigma参数，影响平滑程度和细节保留。
|
|
- Comfy dtype: FLOAT
|
|
- Python dtype: float
|
|
- denoise
|
|
- 决定降噪滤镜的强度。较高的值可以减少噪声，但也可能减弱细节。
|
|
- Comfy dtype: FLOAT
|
|
- Python dtype: float
|
|
- detail_mult
|
|
- 细节增强的乘数。调整这个值可以微调细节增强效果的强度。
|
|
- Comfy dtype: FLOAT
|
|
- Python dtype: float
|
|
|
|
# Output types
|
|
- image
|
|
- 节点的输出，即经过增强处理后的图像。这些图像预期会有更好的细节表现，并可能降低了噪声。
|
|
- Comfy dtype: IMAGE
|
|
- Python dtype: torch.Tensor
|
|
|
|
|
|
## Usage tips
|
|
- Infra type: `GPU`
|
|
- Common nodes: unknown
|
|
|
|
|
|
## Source code
|
|
```python
|
|
class EnhanceDetail:
|
|
def __init__(self):
|
|
pass
|
|
|
|
@classmethod
|
|
def INPUT_TYPES(s):
|
|
return {
|
|
"required": {
|
|
"images": ("IMAGE",),
|
|
"filter_radius": ("INT", {
|
|
"default": 2,
|
|
"min": 1,
|
|
"max": 64,
|
|
"step": 1
|
|
}),
|
|
"sigma": ("FLOAT", {
|
|
"default": 0.1,
|
|
"min": 0.01,
|
|
"max": 100.0,
|
|
"step": 0.01
|
|
}),
|
|
"denoise": ("FLOAT", {
|
|
"default": 0.1,
|
|
"min": 0.0,
|
|
"max": 10.0,
|
|
"step": 0.01
|
|
}),
|
|
"detail_mult": ("FLOAT", {
|
|
"default": 2.0,
|
|
"min": 0.0,
|
|
"max": 100.0,
|
|
"step": 0.1
|
|
}),
|
|
},
|
|
}
|
|
|
|
RETURN_TYPES = ("IMAGE",)
|
|
FUNCTION = "enhance"
|
|
|
|
CATEGORY = "image/filters"
|
|
|
|
def enhance(self, images: torch.Tensor, filter_radius: int, sigma: float, denoise: float, detail_mult: float):
|
|
|
|
if filter_radius == 0:
|
|
return (images,)
|
|
|
|
d = filter_radius * 2 + 1
|
|
s = sigma / 10
|
|
n = denoise / 10
|
|
|
|
dup = copy.deepcopy(images.cpu().numpy())
|
|
|
|
for index, image in enumerate(dup):
|
|
imgB = image
|
|
if denoise>0.0:
|
|
imgB = cv2.bilateralFilter(image, d, n, d)
|
|
|
|
imgG = np.clip(guidedFilter(image, image, d, s), 0.001, 1)
|
|
|
|
details = (imgB/imgG - 1) * detail_mult + 1
|
|
dup[index] = np.clip(details*imgG - imgB + image, 0, 1)
|
|
|
|
return (torch.from_numpy(dup),)
|
|
|
|
```
|