I implemented an algorithm that detects faces and I want to blur faces. I'm using PIL for blurring it.
image = Image.open(path_img)
draw = ImageDraw.Draw(image)
draw.ellipse((top, left, bottom, right), fill = 'white', outline ='white')
I got this with my code
Face to blur
I would like to use :
blurred_image = cropped_image.filter(ImageFilter.GaussianBlur(radius=10 ))
But I can't use it because I'm using an ImageDraw and it works only with Image class. How can I blur with an ellipse (circular) the face?
Thank you
blur the ellipse is the best way
With Pillow, the best way to do this is to use the blurred ellipse as a blending mask for composite.
from PIL import Image, ImageDraw, ImageFilter
def make_ellipse_mask(size, x0, y0, x1, y1, blur_radius):
img = Image.new("L", size, color=0)
draw = ImageDraw.Draw(img)
draw.ellipse((x0, y0, x1, y1), fill=255)
return img.filter(ImageFilter.GaussianBlur(radius=blur_radius))
kitten_image = Image.open("kitten.jpg")
overlay_image = Image.new("RGB", kitten_image.size, color="orange") # This could be a bitmap fill too, but let's just make it orange
mask_image = make_ellipse_mask(kitten_image.size, 150, 70, 350, 250, 5)
masked_image = Image.composite(overlay_image, kitten_image, mask_image)
masked_image.show()
Given this adorable kitten as input, the output is
EDIT: Inspired by Mark Setchell's answer, simply changing the overlay_image line to
overlay_image = kitten_image.filter(ImageFilter.GaussianBlur(radius=15))
gives us this blur variant (with smooth edges for the blur :) )
Not sure if you want to composite something over the image to conceal the contents, or blur it. This is more blurry :-)
Starting with Paddington:
You can go to "Stealth Mode" like this:
#!/usr/bin/env python3
from PIL import Image, ImageDraw, ImageFilter
import numpy as np
# Open image
im = Image.open('paddington.png')
# Make a mask the same size as the image filled with black
mask = Image.new('RGB',im.size)
# Draw a filled white circle onto the black mask
draw = ImageDraw.Draw(mask)
draw.ellipse([90,40,300,250],fill=(255,255,255))
# Blur the entire image
blurred = im.filter(ImageFilter.GaussianBlur(radius=15))
# Select either the original or the blurred image at each pixel, depending on the mask
res = np.where(np.array(mask)>0,np.array(blurred),np.array(im))
# Convert back to PIL Image and save
Image.fromarray(res).save('result.png')
Or, as suggested by #AKX, you can remove the Numpy dependency and make the code a bit smaller too yet still get same result:
#!/usr/bin/env python3
from PIL import Image, ImageDraw, ImageFilter
import numpy as np
# Open image
im = Image.open('paddington.png')
# Make a mask the same size as the image filled with black
mask = Image.new('L',im.size)
# Draw a filled white circle onto the black mask
draw = ImageDraw.Draw(mask)
draw.ellipse([90,40,300,250],fill=255)
# Blur the entire image
blurred = im.filter(ImageFilter.GaussianBlur(radius=15))
# Composite blurred image over sharp one within mask
res = Image.composite(blurred, im, mask)
# Save
res.save('result.png')
Related
i have two images, i need to use numpy and opencv to overlay foreground on top of background using numpy masks.
import numpy as np
import cv2
import matplotlib.pyplot as plt
background = cv2.imread("background.jpg")
foreground = cv2.imread("foreground.png")
foreground = cv2.cvtColor(foreground ,cv2.COLOR_BGR2GRAY)
background = cv2.cvtColor(background ,cv2.COLOR_BGR2GRAY)
arr = []
for i in range(foreground.shape[1]): #parse forground pixels
temp_row = []
for j in range(foreground.shape[0]):
if((foreground [i][j] == 0)):#if pixel transperant draw background
temp_row.append(background[i][j])
else: # draw forground
temp_row.append(foreground[i][j])
arr.append(temp_row)
res_im = np.array(arr)
plt.figure()
plt.imshow(res_im, cmap='gray', vmin=0, vmax=255)
plt.show()
i used this solution but was told i needed to use masks.. help?
Here is one way to do that in Python/Opencv.
Read the transparent foreground image
Read the background image
Extract the alpha channel from the foreground image
Extract the BGR channels from the foreground image
Composite them together using np.where conditional
Save the result
Front:
Back:
import cv2
import numpy as np
# read foreground image
img = cv2.imread('front.png', cv2.IMREAD_UNCHANGED)
# read background image
back = cv2.imread('back.png')
# extract alpha channel from foreground image as mask and make 3 channels
alpha = img[:,:,3]
alpha = cv2.merge([alpha,alpha,alpha])
# extract bgr channels from foreground image
front = img[:,:,0:3]
# blend the two images using the alpha channel as controlling mask
result = np.where(alpha==(0,0,0), back, front)
# save result
cv2.imwrite("front_back.png", result)
# show result
cv2.imshow("RESULT", result)
cv2.waitKey(0)
Result:
I am trying to crop a centered (or not centered) circle from this image:
I stole this code from the existing questions regarding this topic on stack overflow, something goes wrong though:
import cv2
file = 'dog.png'
img = cv2.imread(file)
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
circle = cv2.HoughCircles(img,
3,
dp=1.5,
minDist=10,
minRadius=1,
maxRadius=10)
x = circle[0][0][0]
y = circle[0][0][1]
r = circle[0][0][2]
rectX = (x - r)
rectY = (y - r)
crop_img = img[rectY:(rectY+2*r), rectX:(rectX+2*r)]
cv2.imwrite('dog_circle.png', crop_img)
Output:
Traceback (most recent call last):
File "C:\Users\Artur\Desktop\crop_circle - Kopie\crop_circle.py", line 14, in <module>
x = circle[0][0][0]
TypeError: 'NoneType' object is not subscriptable
cv2.HoughCircles() seems to produce None instead of a cropped circle array. How do I fix this?
first: HoughCircles is used to detect circles on image, not to crop it.
You can't have circle image. Image is always rectangle but some of pixels can be transparent (alpha channel in RGBA) and programs will not display them.
So you can first crop image to have square and later add alpha channel with information which pixels should be visible. And here you can use mask with white circle on black background. At the end you have to save it as png or tiff because jpg can't keep alpha channel.
I use module PIL/pillow for this.
I crop square region in center of image but you can use different coordinates for this.
Next I create grayscale image with the same size and black background and draw white circle/ellipse.
Finally I add this image as alpha channel to cropped image and save it as png.
from PIL import Image, ImageDraw
filename = 'dog.jpg'
# load image
img = Image.open(filename)
# crop image
width, height = img.size
x = (width - height)//2
img_cropped = img.crop((x, 0, x+height, height))
# create grayscale image with white circle (255) on black background (0)
mask = Image.new('L', img_cropped.size)
mask_draw = ImageDraw.Draw(mask)
width, height = img_cropped.size
mask_draw.ellipse((0, 0, width, height), fill=255)
#mask.show()
# add mask as alpha channel
img_cropped.putalpha(mask)
# save as png which keeps alpha channel
img_cropped.save('dog_circle.png')
img_cropped.show()
Result
BTW:
In mask you can use values from 0 to 255 and different pixels may have different transparency - some of them can be half-transparent to make smooth border.
If you want to use it in HTML on own page then you don't have to create circle image because web browser can round corners of image and display it as circle. You have to use CSS for this.
EDIT: Example with more circles on mask.
from PIL import Image, ImageDraw
filename = 'dog.jpg'
# load image
img = Image.open(filename)
# crop image
width, height = img.size
x = (width - height)//2
img_cropped = img.crop((x, 0, x+height, height))
# create grayscale image with white circle (255) on black background (0)
mask = Image.new('L', img_cropped.size)
mask_draw = ImageDraw.Draw(mask)
width, height = img_cropped.size
mask_draw.ellipse((50, 50, width-50, height-50), fill=255)
mask_draw.ellipse((0, 0, 250, 250), fill=255)
mask_draw.ellipse((width-250, 0, width, 250), fill=255)
# add mask as alpha channel
img_cropped.putalpha(mask)
# save as png which keeps alpha channel
img_cropped.save('dog_2.png')
img_cropped.show()
This answer explains how to apply a mask. First, read in the image:
import cv2
import numpy as np
img = cv2.imread('dog.jpg')
Next create a mask, or a blank image that is the same size as the source image:
h,w,_ = img.shape
mask = np.zeros((h,w), np.uint8)
Then, draw a circle on the mask. Change these parameters based on where the face is:
cv2.circle(mask, (678,321), 5, 255, 654)
Finally, mask the source image:
img = cv2.bitwise_and(img, img, mask= mask)
Here is the mask:
And the output:
The idea is to create a black mask then draw the desired region to crop out in white using cv2.circle(). From there we can use cv2.bitwise_and() with the original image and the mask. To crop the result, we can use cv2.boundingRect() on the mask to obtain the ROI then use Numpy slicing to extract the result. For this example I used the center point derived from the image's width and height
import cv2
import numpy as np
# Create mask and draw circle onto mask
image = cv2.imread('1.jpg')
mask = np.zeros(image.shape, dtype=np.uint8)
x,y = image.shape[1], image.shape[0]
cv2.circle(mask, (x//2,y//2), 300, (255,255,255), -1)
# Bitwise-and for ROI
ROI = cv2.bitwise_and(image, mask)
# Crop mask and turn background white
mask = cv2.cvtColor(mask, cv2.COLOR_BGR2GRAY)
x,y,w,h = cv2.boundingRect(mask)
result = ROI[y:y+h,x:x+w]
mask = mask[y:y+h,x:x+w]
result[mask==0] = (255,255,255)
cv2.imshow('result', result)
cv2.waitKey()
I have the following image which has text and a lot of white space underneath the text. I would like to crop the white space such that it looks like the second image.
Cropped Image
Here is what I've done
>>> img = cv2.imread("pg13_gau.jpg.png")
>>> gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
>>> edged = cv2.Canny(gray, 30,300)
>>> (img,cnts, _) = cv2.findContours(edged.copy(), cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
>>> cnts = sorted(cnts, key = cv2.contourArea, reverse = True)[:10]
As many have alluded in the comments, the best way is to invert the image so the black text becomes white, find all the non-zero points in the image then determine what the minimum spanning bounding box would be. You can use this bounding box to finally crop your image. Finding the contours is very expensive and it isn't needed here - especially since your text is axis-aligned. You can use a combination of cv2.findNonZero and cv2.boundingRect to do what you need.
Therefore, something like this would work:
import numpy as np
import cv2
img = cv2.imread('ws.png') # Read in the image and convert to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
gray = 255*(gray < 128).astype(np.uint8) # To invert the text to white
coords = cv2.findNonZero(gray) # Find all non-zero points (text)
x, y, w, h = cv2.boundingRect(coords) # Find minimum spanning bounding box
rect = img[y:y+h, x:x+w] # Crop the image - note we do this on the original image
cv2.imshow("Cropped", rect) # Show it
cv2.waitKey(0)
cv2.destroyAllWindows()
cv2.imwrite("rect.png", rect) # Save the image
The code above exactly lays out what I talked about in the beginning. We read in the image, but we also convert to grayscale as your image is in colour for some reason. The tricky part is the third line of code where I threshold below the intensity of 128 so that the dark text becomes white. This however produces a binary image, so I convert to uint8, then scale by 255. This essentially inverts the text.
Next, given this image we find all of the non-zero coordinates with cv2.findNonZero and we finally put this into cv2.boundingRect which will give you the top-left corner of the bounding box as well as the width and height. We can finally use this to crop the image. Note we do this on the original image and not the inverted one. We use simply NumPy array indexing to do the cropping for us.
Finally, we show the image to show that it works and we save it to disk.
I now get this image:
For the second image, a good thing to do is to remove some of the right border and bottom border. We can do that by cropping the image down to that first. Next, this image contains some very small noisy pixels. I would recommend doing a morphological opening with a very small kernel, then redo the logic we talked about above.
Therefore:
import numpy as np
import cv2
img = cv2.imread('pg13_gau_preview.png') # Read in the image and convert to grayscale
img = img[:-20,:-20] # Perform pre-cropping
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
gray = 255*(gray < 128).astype(np.uint8) # To invert the text to white
gray = cv2.morphologyEx(gray, cv2.MORPH_OPEN, np.ones((2, 2), dtype=np.uint8)) # Perform noise filtering
coords = cv2.findNonZero(gray) # Find all non-zero points (text)
x, y, w, h = cv2.boundingRect(coords) # Find minimum spanning bounding box
rect = img[y:y+h, x:x+w] # Crop the image - note we do this on the original image
cv2.imshow("Cropped", rect) # Show it
cv2.waitKey(0)
cv2.destroyAllWindows()
cv2.imwrite("rect.png", rect) # Save the image
Note: Output image removed due to privacy
Opencv reads the image as a numpy array and it's much simpler to use numpy directly (scikit-image does the same). One possible way of doing it is to read the image as grayscale or convert to it and do the row-wise and column-wise operations as shown in the code snippet below. This will remove the columns and rows when all pixels are of pixel_value (white in this case).
def crop_image(filename, pixel_value=255):
gray = cv2.imread(filename, cv2.IMREAD_GRAYSCALE)
crop_rows = gray[~np.all(gray == pixel_value, axis=1), :]
cropped_image = crop_rows[:, ~np.all(crop_rows == pixel_value, axis=0)]
return cropped_image
and the output:
This would also work:
from PIL import Image, ImageChops
img = Image.open("pUq4x.png")
pixels = img.load()
print (f"original: {img.size[0]} x {img.size[1]}")
xlist = []
ylist = []
for y in range(0, img.size[1]):
for x in range(0, img.size[0]):
if pixels[x, y] != (255, 255, 255, 255):
xlist.append(x)
ylist.append(y)
left = min(xlist)
right = max(xlist)
top = min(ylist)
bottom = max(ylist)
img = img.crop((left-10, top-10, right+10, bottom+10))
img.show()
I performed a foreground extraction using grabcut on an image. Now, I want to calculate the area of the resulted foreground portion of image.
One way i thought i could do this is by thresholding image, that is foreground becomes white and background becomes black. And then by calculating the number of white pixels -
import cv2
import numpy as np
from matplotlib import pyplot as plt
img = cv2.imread('image.jpg')
mask = np.zeros(img.shape[:2],np.uint8)
bgdModel = np.zeros((1,65),np.float64)
fgdModel = np.zeros((1,65),np.float64)
rect = (1200,1050,1950,1250)
#applying grabcut
cv2.grabCut(img,mask,rect,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_RECT)
mask2 = np.where((mask==2)|(mask==0),0,1).astype('uint8')
img = img*mask2[:,:,np.newaxis]
#thresholding
cvimg = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
ret,thresh = cv2.threshold(cvimg,0,255,cv2.THRESH_BINARY)
count = cv2.countNonZero(thresh)
print("White " , count)
cv2.imshow("image",thresh)
plt.imshow(img)
plt.colorbar()
plt.show()
But i doubt the accuracy as no. of pixels change when resolution of image changes. How can i find the area of foreground of images with a common unit of measurement for all images? Is there any other simple and appropriate method exists?
Thanks for helping...
How would I take an RGB image in Python and convert it to black and white? Not grayscale, I want each pixel to be either fully black (0, 0, 0) or fully white (255, 255, 255).
Is there any built-in functionality for getting it done in the popular Python image processing libraries? If not, would the best way be just to loop through each pixel, if it's closer to white set it to white, if it's closer to black set it to black?
Scaling to Black and White
Convert to grayscale and then scale to white or black (whichever is closest).
Original:
Result:
Pure Pillow implementation
Install pillow if you haven't already:
$ pip install pillow
Pillow (or PIL) can help you work with images effectively.
from PIL import Image
col = Image.open("cat-tied-icon.png")
gray = col.convert('L')
bw = gray.point(lambda x: 0 if x<128 else 255, '1')
bw.save("result_bw.png")
Alternatively, you can use Pillow with numpy.
Pillow + Numpy Bitmasks Approach
You'll need to install numpy:
$ pip install numpy
Numpy needs a copy of the array to operate on, but the result is the same.
from PIL import Image
import numpy as np
col = Image.open("cat-tied-icon.png")
gray = col.convert('L')
# Let numpy do the heavy lifting for converting pixels to pure black or white
bw = np.asarray(gray).copy()
# Pixel range is 0...255, 256/2 = 128
bw[bw < 128] = 0 # Black
bw[bw >= 128] = 255 # White
# Now we put it back in Pillow/PIL land
imfile = Image.fromarray(bw)
imfile.save("result_bw.png")
Black and White using Pillow, with dithering
Using pillow you can convert it directly to black and white. It will look like it has shades of grey but your brain is tricking you! (Black and white near each other look like grey)
from PIL import Image
image_file = Image.open("cat-tied-icon.png") # open colour image
image_file = image_file.convert('1') # convert image to black and white
image_file.save('/tmp/result.png')
Original:
Converted:
Black and White using Pillow, without dithering
from PIL import Image
image_file = Image.open("cat-tied-icon.png") # open color image
image_file = image_file.convert('1', dither=Image.NONE) # convert image to black and white
image_file.save('/tmp/result.png')
I would suggest converting to grayscale, then simply applying a threshold (halfway, or mean or meadian, if you so choose) to it.
from PIL import Image
col = Image.open('myimage.jpg')
gry = col.convert('L')
grarray = np.asarray(gry)
bw = (grarray > grarray.mean())*255
imshow(bw)
img_rgb = cv2.imread('image.jpg')
img_gray = cv2.cvtColor(img_rgb, cv2.COLOR_BGR2GRAY)
(threshi, img_bw) = cv2.threshold(img_gray, 0, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)
Pillow, with dithering
Using pillow you can convert it directly to black and white. It will look like it has shades of grey but your brain is tricking you! (Black and white near each other look like grey)
from PIL import Image
image_file = Image.open("cat-tied-icon.png") # open colour image
image_file = image_file.convert('1') # convert image to black and white
image_file.save('/tmp/result.png')
Original:
Converted:
And you can use colorsys (in the standard library) to convert rgb to hls and use the lightness value to determine black/white:
import colorsys
# convert rgb values from 0-255 to %
r = 120/255.0
g = 29/255.0
b = 200/255.0
h, l, s = colorsys.rgb_to_hls(r, g, b)
if l >= .5:
# color is lighter
result_rgb = (255, 255, 255)
elif l < .5:
# color is darker
result_rgb = (0,0,0)
Using opencv You can easily convert rgb to binary image
import cv2
%matplotlib inline
import matplotlib.pyplot as plt
from skimage import io
from PIL import Image
import numpy as np
img = io.imread('http://www.bogotobogo.com/Matlab/images/MATLAB_DEMO_IMAGES/football.jpg')
img = cv2.cvtColor(img, cv2.IMREAD_COLOR)
imR=img[:,:,0] #only taking gray channel
print(img.shape)
plt.imshow(imR, cmap=plt.get_cmap('gray'))
#Gray Image
plt.imshow(imR)
plt.title('my picture')
plt.show()
#Histogram Analyze
imgg=imR
hist = cv2.calcHist([imgg],[0],None,[256],[0,256])
plt.hist(imgg.ravel(),256,[0,256])
# show the plotting graph of an image
plt.show()
#Black And White
height,width=imgg.shape
for i in range(0,height):
for j in range(0,width):
if(imgg[i][j]>60):
imgg[i][j]=255
else:
imgg[i][j]=0
plt.imshow(imgg)
Here is the code for creating binary image using opencv-python :
img = cv2.imread('in.jpg',2)
ret, bw_img = cv2.threshold(img,127,255,cv2.THRESH_BINARY)
cv2.imshow("Output - Binary Image",bw_img)
If you don't want to use cv methods for the segmentation and understand what you are doing, treat the RGB image as matrix.
image = mpimg.imread('image_example.png') # your image
R,G,B = image[:,:,0], image[:,:,1], image[:,:,2] # the 3 RGB channels
thresh = [100, 200, 50] # example of triple threshold
# First, create an array of 0's as default value
binary_output = np.zeros_like(R)
# then screen all pixels and change the array based on RGB threshold.
binary_output[(R < thresh[0]) & (G > thresh[1]) & (B < thresh[2])] = 255
The result is an array of 0's and 255's based on a triple condition.