We Keep Coding

Convert polygons (in the image) to masks

I have some images and their associated ground truth outlined objects. For example this image shows the outlined objects for one of the original imagesoutlined objects in blue
Given this image and its original source, I would like to create some masks based on these outlines using openCV2 or skimage.
Using Contours I can roughly achieve that, but I have two problems:
1- Why I get repeated masks? (plz refer to the attached snippet)
2- How to overcome the issue of two touching objects
from skimage import io
from skimage import measure
import matplotlib.pyplot as plt
image = io.imread('path/to/the/attached/image')
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
contours = measure.find_contours(gray, 0.1)
for n, contour in enumerate(contours):
r_mask = np.zeros_like(gray, dtype='bool')
r_mask[np.round(contour[:, 0]).astype('int'), np.round(contour[:,
1]).astype('int')] = 1
r_mask = ndimage.binary_fill_holes(r_mask)
io.imshow(r_mask)
plt.show()
Thank you

Finding bright spots in a image using opencv

I want to find the bright spots in the above image and tag them using some symbol. For this i have tried using the Hough Circle Transform algorithm that OpenCV already provides. But it is giving some kind of assertion error when i run the code. I also tried the Canny edge detection algorithm which is also provided in OpenCV but it is also giving some kind of assertion error. I would like to know if there is some method to get this done or if i can prevent those error messages.
I am new to OpenCV and any help would be really appreciated.
P.S. - I can also use Scikit-image if necessary. So if this can be done using Scikit-image then please tell me how.
Below is my preprocessing code:
import cv2
import numpy as np
image = cv2.imread("image1.png")
gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
binary_image = np.where(gray_image > np.mean(gray_image),1.0,0.0)
binary_image = cv2.Laplacian(binary_image, cv2.CV_8UC1)

If you are just going to work with simple images like your example where you have black background, you can use same basic preprocessing/thresholding then find connected components. Use this example code to draw a circle inside all circles in the image.
import cv2
import numpy as np
image = cv2.imread("image1.png")
# constants
BINARY_THRESHOLD = 20
CONNECTIVITY = 4
DRAW_CIRCLE_RADIUS = 4
# convert to gray
gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# extract edges
binary_image = cv2.Laplacian(gray_image, cv2.CV_8UC1)
# fill in the holes between edges with dilation
dilated_image = cv2.dilate(binary_image, np.ones((5, 5)))
# threshold the black/ non-black areas
_, thresh = cv2.threshold(dilated_image, BINARY_THRESHOLD, 255, cv2.THRESH_BINARY)
# find connected components
components = cv2.connectedComponentsWithStats(thresh, CONNECTIVITY, cv2.CV_32S)
# draw circles around center of components
#see connectedComponentsWithStats function for attributes of components variable
centers = components[3]
for center in centers:
cv2.circle(thresh, (int(center[0]), int(center[1])), DRAW_CIRCLE_RADIUS, (255), thickness=-1)
cv2.imwrite("res.png", thresh)
cv2.imshow("result", thresh)
cv2.waitKey(0)
Here is resulting image:
Edit: connectedComponentsWithStats takes a binary image as input, and returns connected pixel groups in that image. If you would like to implement that function yourself, naive way would be:
1- Scan image pixels from top left to bottom right until you encounter a non-zero pixel that does not have a label (id).
2- When you encounter a non-zero pixel, search all its neighbours recursively( If you use 4 connectivity you check UP-LEFT-DOWN-RIGHT, with 8 connectivity you also check diagonals) until you finish that region. Assign each pixel a label. Increase your label counter.
3- Continue scanning from where you left.

Draw circle around imperfect circular objects

I have this image of an eye where I want to get the center of the pupil:
Original Image
I applied adaptive threshold as well as laplacian to the image using this code:
import cv2
import numpy as np
from matplotlib import pyplot as plt
img = cv2.imread('C:\Users\User\Documents\module4\input\left.jpg',0)
image = cv2.medianBlur(img,5)
th = cv2.adaptiveThreshold(image,255,cv2.ADAPTIVE_THRESH_MEAN_C,
cv2.THRESH_BINARY,11,2)
laplacian = cv2.Laplacian(th,cv2.CV_64F)
cv2.imshow('output', laplacian)
cv2.imwrite('C:\Users\User\Documents\module4\output\output.jpg', laplacian)
cv2.waitKey(0)
cv2.destroyAllWindows
and the resulting image looks like this: Resulting image by applying adaptive threshold
I want to draw a circle around the smaller inner circle and get its center. I've tried using contours and circular hough transform but it does not correctly detect any circles in the image.
Here is my code for Circular Hough Transform:
import cv2
import numpy as np
from matplotlib import pyplot as plt
img = cv2.imread('C:\Users\User\Documents\module4\output\output.jpg',0)
circles = cv2.HoughCircles(img,cv2.HOUGH_GRADIENT,1,20,param1=50,param2=30,minRadius=0,maxRadius=0)
circles = np.uint16(np.around(circles))
for i in circles[0,:]:
# draw the outer circle
cv2.circle(img,(i[0],i[1]),i[2],(255,255,0),2)
# draw the center of the circle
cv2.circle(img,(i[0],i[1]),2,(255,0,255),3)
cv2.imshow('detected circles',img)
cv2.waitKey(0)
cv2.destroyAllWindows()
And here is the code for applying contour:
import cv2
import numpy as np
img = cv2.imread('C:\Users\User\Documents\module4\output\output.jpg',0)
_, contours,hierarchy = cv2.findContours(img, 1, 2)
cnt = contours[0]
(x,y),radius = cv2.minEnclosingCircle(cnt)
center = (int(x),int(y))
radius = int(radius)
img = cv2.circle(img,center,radius,(0,255,255),2)
cv2.imshow('contour', img)
cv2.waitKey(0)
cv2.destroyAllWindows()
The resulting image of this code exactly looks like the image wherein I applied adaptive threshold. I would really appreciate it if anyone can help me solve my problem. I've been stuck with this for a while now. Also, if any of you guys can suggest a better way to detect the center of the pupil besides this method, I would also really appreciate it.

try to apply edge detection instead of shareholding after filtering of original image and then apply hough circle

My thought would be to use the Hough transform like you're doing. But another method might be template matching like this. This assumes you know the approximate radius of the pupil in the image, you can try to build a template.
import skimage
import numpy as np
import matplotlib.pyplot as plt
img = skimage.io.imread('Wjioe.jpg')
#just use grayscale, but you could make separate template for each r,g,b channel
img = np.mean(img, axis=2)
(M,N) = img.shape
mm = M-20
nn = N-20
template = np.zeros([mm,nn])
## Create template ##
#darkest inner circle (pupil)
(rr,cc) = skimage.draw.circle(mm/2,nn/2,4.5, shape=template.shape)
template[rr,cc]=-2
#iris (circle surrounding pupil)
(rr,cc) = skimage.draw.circle(mm/2,nn/2,8, shape=template.shape)
template[rr,cc] = -1
#Optional - pupil reflective spot (if centered)
(rr,cc) = skimage.draw.circle(mm/2,nn/2,1.5, shape=template.shape)
template[rr,cc] = 1
plt.imshow(template)
normccf = skimage.feature.match_template(img, template,pad_input=True)
#center pixel
(i,j) = np.unravel_index( np.argmax(normccf), normccf.shape)
plt.imshow(img)
plt.plot(j,i,'r*')

You're defining a 3 channel color for a gray-scale image. Based on my test it will only read the first value in that tuple. Because the first value in your other colors (in the middle code) starts with 255, it draws a full white circle and because the first value in your last color (in your last code) starts with 0, it draws a full black circle which you can't see.
Just change your color values to a 1 channel color (an int between 0 and 255) and you'll be fine.

Filling "holes" of an image in python with cv2 not working

I am trying to fill the "holes" of red blood cells in an image after performing binary threshold. Almost all red blood cells have a black center when inverting the binary threshold. I want to remove them.
Example image:
This is my code:
import cv2
from PIL import Image
import numpy as np
from scipy import ndimage
from skimage.feature import peak_local_max
from skimage.morphology import watershed
image = cv2.imread("blood_cells.jpg")
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
darker = cv2.equalizeHist(gray)
ret,thresh = cv2.threshold(darker,0,255,cv2.THRESH_BINARY+cv2.THRESH_OTSU)
newimg = cv2.bitwise_not(thresh)
im2, contours, hierarchy = cv2.findContours(newimg,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
for cnt in contours:
cv2.drawContours(newimg,[cnt],0,255,-1)
And it worked. I filled the holes using findContours() and drawContours().
but when I try to compute the euclidean distance, for applying the watershed algorithm, I get only 52 unique segments, however there should be more. Here is the code, if it might be helpful:
D = ndimage.distance_transform_edt(newimg)
localMax = peak_local_max(D, indices=False, min_distance=20, labels=thresh)
markers = ndimage.label(localMax, structure=np.ones((3, 3)))[0]
labels = watershed(-D, markers, mask=thresh)
print("[INFO] {} unique segments found".format(len(np.unique(labels)) - 1))
I tried to segment each cell, but the results were quite off. Only the inside of the cells that had "holes" got segmented.
First image shows my result, second shows how it should roughly look like:
.
I then filled the holes manually, just to see if my code for segmentation works - and it works. The error should be somewhere between the part where I drew the contours, and the part where I calculated the euclidean distance.. Could anybody explain to me what could be wrong? I am clueless.

Your problem lies in the following line:
labels = watershed(-D, markers, mask=thresh)
You're passing as mask an inverted, uncorrected result from thresholding:
Giving you this bad segmentation:
Whereas you should be passing the corrected, filled in mask:
labels = watershed(-D, markers, mask=newimg)
Giving you the result you probably expect:

Remove features from binarized image

I wrote a little script to transform pictures of chalkboards into a form that I can print off and mark up.
I take an image like this:
Auto-crop it, and binarize it. Here's the output of the script:
I would like to remove the largest connected black regions from the image. Is there a simple way to do this?
I was thinking of eroding the image to eliminate the text and then subtracting the eroded image from the original binarized image, but I can't help thinking that there's a more appropriate method.

Sure you can just get connected components (of certain size) with findContours or floodFill, and erase them leaving some smear. However, if you like to do it right you would think about why do you have the black area in the first place.
You did not use adaptive thresholding (locally adaptive) and this made your output sensitive to shading. Try not to get the black region in the first place by running something like this:
Mat img = imread("desk.jpg", 0);
Mat img2, dst;
pyrDown(img, img2);
adaptiveThreshold(255-img2, dst, 255, ADAPTIVE_THRESH_MEAN_C,
THRESH_BINARY, 9, 10); imwrite("adaptiveT.png", dst);
imshow("dst", dst);
waitKey(-1);
In the future, you may read something about adaptive thresholds and how to sample colors locally. I personally found it useful to sample binary colors orthogonally to the image gradient (that is on the both sides of it). This way the samples of white and black are of equal size which is a big deal since typically there are more background color which biases estimation. Using SWT and MSER may give you even more ideas about text segmentation.

I tried this:
import numpy as np
import cv2
im = cv2.imread('image.png')
gray = cv2.cvtColor(im,cv2.COLOR_BGR2GRAY)
grayout = 255*np.ones((im.shape[0],im.shape[1],1), np.uint8)
blur = cv2.GaussianBlur(gray,(5,5),1)
thresh = cv2.adaptiveThreshold(blur,255,1,1,11,2)
contours,hierarchy = cv2.findContours(thresh,cv2.RETR_LIST,cv2.CHAIN_APPROX_SIMPLE)
wcnt = 0
for item in contours:
area =cv2.contourArea(item)
print wcnt,area
[x,y,w,h] = cv2.boundingRect(item)
if area>10 and area<200:
roi = gray[y:y+h,x:x+w]
cntd = 0
for i in range(x,x+w):
for j in range(y,y+h):
if gray[j,i]==0:
cntd = cntd + 1
density = cntd/(float(h*w))
if density<0.5:
for i in range(x,x+w):
for j in range(y,y+h):
grayout[j,i] = gray[j,i];
wcnt = wcnt + 1
cv2.imwrite('result.png',grayout)
You have to balance two things, removing the black spots but balance that with not losing the contents of what is on the board. The output I got is this:

Here is a Python numpy implementation (using my own mahotas package) of the method for the top answer (almost the same, I think):
import mahotas as mh
import numpy as np
Imported mahotas & numpy with standard abbreviations
im = mh.imread('7Esco.jpg', as_grey=1)
Load the image & convert to gray
im2 = im[::2,::2]
im2 = mh.gaussian_filter(im2, 1.4)
Downsample and blur (for speed and noise removal).
im2 = 255 - im2
Invert the image
mean_filtered = mh.convolve(im2.astype(float), np.ones((9,9))/81.)
Mean filtering is implemented "by hand" with a convolution.
imc = im2 > mean_filtered - 4
You might need to adjust the number 4 here, but it worked well for this image.
mh.imsave('binarized.png', (imc*255).astype(np.uint8))
Convert to 8 bits and save in PNG format.