使用分水岭，skikit图像计数细胞数和细胞核数

Question

我正在尝试量化大量的图像数据。每个图像都有细胞和细胞核。 必须要做的是以示意图的形式表示为“我需要的东西”：

示例图像显示在“原始图像”中：

我找到了一个在线分水岭算法程序来计算细胞，但我无法计算细胞内部（和外部）的细胞核数量。

这是我用来计算原始图像中的单元格的程序

#import packages

import numpy as np
import matplotlib.pyplot as plt
from skimage import io, color, filters as filters
from scipy import ndimage
from skimage.morphology import watershed
from skimage.feature import peak_local_max
from skimage.measure import regionprops, label
import numpy as np
from scipy.ndimage import gaussian_filter
import matplotlib.pyplot as plt
from skimage import data
from skimage import img_as_float
from skimage.morphology import reconstruction
import skimage
from skimage import segmentation

%matplotlib inline
import matplotlib
#import image
from IPython.core.display import Image

Image(filename=('/Users/sasi/Desktop/image1.jpeg'))

# Find number of cells

image = color.rgb2gray(io.imread('/Users/sasi/Desktop/image1.jpeg'))
image = image < filters.threshold_otsu(image)
distance = ndimage.distance_transform_edt(image)
local_maxi = peak_local_max(distance, indices=False, footprint=np.ones((3, 3)), labels=image)

markers, num_features = ndimage.label(local_maxi)
labels = watershed(-distance, markers, mask=image)
regions = regionprops(labels)
regions = [r for r in regions if r.area > 60]

print('Number of cells:', len(regions) - 1)

我应该如何计算细胞内的细胞核以及整个图像中的核数？另外，如果你有另一个更好的计算细胞计划，请告诉我。

Answer 1

    import cv2
    import numpy as np
    from skimage.feature import peak_local_max
    from skimage.morphology import watershed
    from scipy import ndimage

    image=cv2.imread("Da0003.jpg")
    gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    sigma=2
    gauss_img = cv2.GaussianBlur(gray_image,(0,0),1.6*sigma,0)
    main_img = cv2.GaussianBlur(gray_image,(0,0),sigma,0)
    ret,threshold_img = cv2.threshold(main_img,127,255,cv2.THRESH_BINARY)
    threshold=cv2.bitwise_not(threshold_img)
    D = ndimage.distance_transform_edt(threshold)
    localMax = peak_local_max(D, indices=False, min_distance=10,
        labels=threshold)

    markers = ndimage.label(localMax, structure=np.ones((3, 3)))[0]
    labels = watershed(-D, markers, mask=threshold)
    for label in np.unique(labels):
        # if the label is zero, we are examining the 'background'
        # so simply ignore it
            if label == 0:
                continue

            # otherwise, allocate memory for the label region and draw
            # it on the mask
            x,y,_=image.shape
            image_copy=image  
            mask = np.zeros((x,y), dtype="uint8")
            mask[labels == label] = 255


            cnts = cv2.findContours(mask.copy(), cv2.RETR_EXTERNAL,
                cv2.CHAIN_APPROX_SIMPLE)[-2]
            c = max(cnts, key=cv2.contourArea)

            cv2.drawContours(image_copy, [c], -1, (0, 255, 255), 1)
            #print(label)

    cv2.imwrite("_watershead"+".jpg", image_copy)
    print("count:"+str(label))