update readme

.gitignore

@@ -1,2 +1,3 @@
+_test_*.*
 __pycache__
 model.pth

README.MD

@@ -13,7 +13,7 @@ Nodes are divided into 5 groups according to their functions: LayerStyle, LayerC
 [中文说明点这里](./README_CN.MD)
 
 ## Update
-**Due to the new nodes, if error occurs during runtime, please reinstall the dependency package.
+**If the dependency package error after updating, please reinstall the relevant dependency packages. for details, please refer to [here](https://github.com/chflame163/ComfyUI_LayerStyle/issues/5).
 * Commit [Sharp & Soft](#Sharp) node, it can enhance or smooth out image details. Commit [MaskByDifferent](#MaskByDifferent) node, it compare two images and output a Mask. Commit [SegmentAnythingUltra](#SegmentAnythingUltra) node, Improve the quality of mask edges. *If SegmentAnything is not installed, you will need to manually download the model.
 * All nodes have fully supported batch images, providing convenience for video creation.
 (The CropByMask node only supports cuts of the same size. if a batch mask_for_crop inputted, the data from the first sheet will be used.)

README_CN.MD

@@ -10,7 +10,7 @@
 * [LayerFilter](#LayerFilter)节点组提供图像效果滤镜。
 
 ## 更新说明
-**由于引入新的节点，请重新安装依赖包。
+**如果本插件更新后出现依赖包错误，请重新安装相关依赖包。详情见[这里](https://github.com/chflame163/ComfyUI_LayerStyle/issues/5)。
 * 新增[Sharp & Soft](#Sharp)节点，可提升或抹平图像细节。新增[MaskByDifferent](#MaskByDifferent)节点，比较两张图片并输出Mask。新增[SegmentAnythingUltra](#SegmentAnythingUltra)节点，提升遮罩边缘质量。*如果没有安装SegmentAnything, 需要手动下载模型。
 * 所有节点已全面支持批量图片，为创作视频提供方便。( CropByMask 节点仅支持相同尺寸的切除, 如果输入批量mask_for_crop，将使用第一张的数据。)
 * 添加[RemBgUltra](#RemBgUltra) 和 [PixelSpread](#PixelSpread) 节点，显著提升了遮罩质量。*RemBgUltra需手动下载模型。

py/imagefunc.py

@@ -16,17 +16,24 @@ import time
 from typing import Union, List
 from PIL import Image, ImageFilter, ImageChops, ImageDraw, ImageOps, ImageEnhance, ImageFont
 from skimage import img_as_float, img_as_ubyte
-from pymatting import fix_trimap, estimate_alpha_cf
+from pymatting import fix_trimap, estimate_alpha_cf, estimate_foreground_ml
 import torchvision.transforms.functional as TF
 import torch.nn.functional as F
 import colorsys
 from .briarmbg import BriaRMBG
 
+try:
+    from cv2.ximgproc import guidedFilter
+except ImportError:
+    print(f'# 😺dzNodes: \033[33mDependency package error -> Unable import "guidedFilter", please reinstall "opencv-contrib-python"\033[m')
+
 current_directory = os.path.dirname(os.path.abspath(__file__))
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
-def log(message):
+def log(message:str, message_type:str='info'):
     name = 'LayerStyle'
+    if message_type == 'error':
+        message = '\033[33m' + message + '\033[m'
     print(f"# 😺dzNodes: {name} -> {message}")
 
 '''Converter'''
@@ -668,8 +675,7 @@ def RMBG(image:Image) -> Image:
     return _mask
 
 def mask_edge_detail(image:torch.Tensor, mask:Image, detail_range:int=8, black_point:float=0.01, white_point:float=0.99) -> torch.Tensor:
-
-    d = detail_range * 2 + 1
+    d = detail_range * 5 + 1
     i_dup = copy.deepcopy(image.cpu().numpy().astype(np.float64))
     a_dup = copy.deepcopy(pil2tensor(mask.convert('RGB')).cpu().numpy().astype(np.float64))
     for index, img in enumerate(i_dup):
@@ -682,6 +688,15 @@ def mask_edge_detail(image:torch.Tensor, mask:Image, detail_range:int=8, black_p
         a_dup[index] = np.stack([alpha, alpha, alpha], axis=-1)  # convert back to rgb
     return torch.from_numpy(a_dup.astype(np.float32))
 
+def guided_filter_alpha(image:torch.Tensor, mask:Image, filter_radius:int, sigma:float) -> torch.Tensor:
+    d = filter_radius + 1
+    s = sigma / 10
+    i_dup = copy.deepcopy(image.cpu().numpy())
+    a_dup = copy.deepcopy(pil2tensor(mask.convert('RGB')).cpu().numpy())
+    for index, image in enumerate(i_dup):
+        alpha_work = a_dup[index]
+        i_dup[index] = guidedFilter(image, alpha_work, d, s)
+    return torch.from_numpy(i_dup)
 
 def mask_fix(images:torch.Tensor, radius:int, fill_holes:int, white_threshold:float, extra_clip:float) -> torch.Tensor:
     d = radius * 2 + 1
@@ -707,6 +722,13 @@ def mask_fix(images:torch.Tensor, radius:int, fill_holes:int, white_threshold:fl
         i_dup[index] = cleaned
     return torch.from_numpy(i_dup)
 
+def histogram_remap(image:torch.Tensor, blackpoint:float, whitepoint:float) -> torch.Tensor:
+    bp = min(blackpoint, whitepoint - 0.001)
+    scale = 1 / (whitepoint - bp)
+    i_dup = copy.deepcopy(image.cpu().numpy())
+    i_dup = np.clip((i_dup - bp) * scale, 0.0, 1.0)
+    return torch.from_numpy(i_dup)
+
 def expand_mask(mask:torch.Tensor, grow:int, blur:int) -> torch.Tensor:
     # grow
     c = 0

py/mask_by_different.py

@@ -15,9 +15,9 @@ class MaskByDifferent:
                 "image_1": ("IMAGE", ),  #
                 "image_2": ("IMAGE",),  #
                 "gain": ("FLOAT", {"default": 1.5, "min": 0.1, "max": 100, "step": 0.1}),
-                "fix_gap": ("INT", {"default": 4, "min": 0, "max": 16, "step": 1}),
-                "fix_threshold": ("FLOAT", {"default": 0.75, "min": 0.01, "max": 1.0, "step": 0.01}),
-                "main_subject_detect": ("BOOLEAN", {"default": True}),
+                "fix_gap": ("INT", {"default": 4, "min": 0, "max": 32, "step": 1}),
+                "fix_threshold": ("FLOAT", {"default": 0.75, "min": 0.01, "max": 0.99, "step": 0.01}),
+                "main_subject_detect": ("BOOLEAN", {"default": False}),
             },
             "optional": {
             }

py/pixel_spread.py

@@ -1,4 +1,3 @@
-from pymatting import estimate_foreground_ml
 from .imagefunc import *
 
 NODE_NAME = 'PixelSpread'

py/segment_anything_ultra.py

@@ -1,5 +1,3 @@
-import torch
-
 from .imagefunc import *
 from .segment_anything_func import *
 

py/sharp&soft.py

@@ -1,4 +1,3 @@
-from cv2.ximgproc import guidedFilter
 from .imagefunc import *
 
 NODE_NAME = 'Sharp & Soft'

requirements.txt

@@ -3,7 +3,6 @@ pillow
 torch
 matplotlib
 Scipy
-opencv-python
 scikit_image
 opencv-contrib-python
 pymatting