In this repo will show the ability of image preprocessing to detect and mark cell of human tissue.
The two below images are an example of image tissue and with which will work with.
def image_preprocessing(given_image):
image_gray_scale = convert_to_gray_scale(given_image)
_, image_threshold_bw = cv2.threshold(image_gray_scale, 100, 255, cv2.THRESH_OTSU)
kernel = np.ones((3, 3), np.uint8)
image_morph = cv2.morphologyEx(image_threshold_bw, cv2.MORPH_CLOSE, kernel, iterations=3)
image_mask = 255 - image_morph
background = cv2.dilate(image_mask, kernel, iterations=2)
dist_transform = cv2.distanceTransform(image_mask, cv2.DIST_L2, 3)
_, foreground = cv2.threshold(dist_transform, 0.285 * dist_transform.max(), 255, 0)
the_unknown_image = background - foreground
foreground = np.uint8(foreground)def markers_creation(foreground, the_unknown_image):
_, image_markers = cv2.connectedComponents(foreground, connectivity=8)
image_markers = image_markers + 10
image_markers[the_unknown_image == 255] = 0
show_image_with_matplot(image_markers)
return image_markersIn the study of image processing, a watershed is a transformation defined on a grayscale image. The name refers metaphorically to a geological watershed, or drainage divide, which separates adjacent drainage basins. The watershed transformation treats the image it operates upon like a topographic map, with the brightness of each point representing its height, and finds the lines that run along the tops of ridges. link
def watershed(pure_image, image_markers):
image_markers = cv2.watershed(pure_image, image_markers)
pure_image[image_markers == -1] = [0, 255, 0]
image_label2rgb = color.label2rgb(image_markers, bg_label=0)
plot_an_image(pure_image)
Using the second image to apply the same parameters and filters to observ the results.
In order to improve the accuracy detection better into the cells. It will be wise to tunne the filters and thresholding parameters and reach a better result.