88 lines
3.6 KiB
Markdown
88 lines
3.6 KiB
Markdown
# Documentation
|
|
- Class name: PrepImageForClipVision
|
|
- Category: ipadapter/utils
|
|
- Output node: False
|
|
- Repo Ref: https://github.com/cubiq/ComfyUI_IPAdapter_plus.git
|
|
|
|
该节点旨在为ClipVision模型预处理图像，确保它们被正确格式化以供分析。它专注于调整图像大小、裁剪和锐化图像，以增强其特征并符合模型的输入要求。
|
|
|
|
# Input types
|
|
## Required
|
|
- image
|
|
- 图像参数至关重要，因为它是节点处理的主要输入。它影响节点的整体操作，决定了输出的质量和特征。
|
|
- Comfy dtype: IMAGE
|
|
- Python dtype: PIL.Image or torch.Tensor
|
|
## Optional
|
|
- interpolation
|
|
- 插值模式定义了应该如何调整图像大小，这会显著影响结果图像的清晰度和细节。这是一个可选参数，允许在调整大小后控制图像的外观。
|
|
- Comfy dtype: COMBO[('LANCZOS', 'BICUBIC', 'HAMMING', 'BILINEAR', 'BOX', 'NEAREST')]
|
|
- Python dtype: str
|
|
- crop_position
|
|
- 裁剪位置决定了调整大小后图像的裁剪方式。这影响了图像的构图和焦点，确保最相关的特征被居中或适当定位。
|
|
- Comfy dtype: COMBO[('top', 'bottom', 'left', 'right', 'center', 'pad')]
|
|
- Python dtype: str
|
|
- sharpening
|
|
- 锐化调整图像的对比度，增强其边缘和细节。该参数允许微调图像的视觉清晰度，并且可以显著提高模型识别特征的能力。
|
|
- Comfy dtype: FLOAT
|
|
- Python dtype: float
|
|
|
|
# Output types
|
|
- output
|
|
- 输出是处理后的图像，现在已格式化并准备好供ClipVision模型分析。它代表了节点处理的结晶，包含了对原始图像进行的所有调整。
|
|
- Comfy dtype: IMAGE
|
|
- Python dtype: torch.Tensor
|
|
|
|
# Usage tips
|
|
- Infra type: CPU
|
|
|
|
# Source code
|
|
```
|
|
class PrepImageForClipVision:
|
|
|
|
@classmethod
|
|
def INPUT_TYPES(s):
|
|
return {'required': {'image': ('IMAGE',), 'interpolation': (['LANCZOS', 'BICUBIC', 'HAMMING', 'BILINEAR', 'BOX', 'NEAREST'],), 'crop_position': (['top', 'bottom', 'left', 'right', 'center', 'pad'],), 'sharpening': ('FLOAT', {'default': 0.0, 'min': 0, 'max': 1, 'step': 0.05})}}
|
|
RETURN_TYPES = ('IMAGE',)
|
|
FUNCTION = 'prep_image'
|
|
CATEGORY = 'ipadapter/utils'
|
|
|
|
def prep_image(self, image, interpolation='LANCZOS', crop_position='center', sharpening=0.0):
|
|
size = (224, 224)
|
|
(_, oh, ow, _) = image.shape
|
|
output = image.permute([0, 3, 1, 2])
|
|
if crop_position == 'pad':
|
|
if oh != ow:
|
|
if oh > ow:
|
|
pad = (oh - ow) // 2
|
|
pad = (pad, 0, pad, 0)
|
|
elif ow > oh:
|
|
pad = (ow - oh) // 2
|
|
pad = (0, pad, 0, pad)
|
|
output = T.functional.pad(output, pad, fill=0)
|
|
else:
|
|
crop_size = min(oh, ow)
|
|
x = (ow - crop_size) // 2
|
|
y = (oh - crop_size) // 2
|
|
if 'top' in crop_position:
|
|
y = 0
|
|
elif 'bottom' in crop_position:
|
|
y = oh - crop_size
|
|
elif 'left' in crop_position:
|
|
x = 0
|
|
elif 'right' in crop_position:
|
|
x = ow - crop_size
|
|
x2 = x + crop_size
|
|
y2 = y + crop_size
|
|
output = output[:, :, y:y2, x:x2]
|
|
imgs = []
|
|
for img in output:
|
|
img = T.ToPILImage()(img)
|
|
img = img.resize(size, resample=Image.Resampling[interpolation])
|
|
imgs.append(T.ToTensor()(img))
|
|
output = torch.stack(imgs, dim=0)
|
|
del imgs, img
|
|
if sharpening > 0:
|
|
output = contrast_adaptive_sharpening(output, sharpening)
|
|
output = output.permute([0, 2, 3, 1])
|
|
return (output,)
|
|
``` |