Python script to generate dotted text from any image

Prerequisites: Python OpenCV

In this article, we will see how we can use Python3 to create a dotted text that represents an image. We use such type of texts in messages and social media and generating them is an interesting image processing task.

This task will need us to define the edges of the given image. After that we can use any symbol like ‘.’, ‘o’ or ‘x’ to display the bright and dark pixels. 

• First create a new python script and install open CV.

python3 -m pip install opencv-python --user

• Now import cv2 and load the image you want to convert. Here the image name is “mic.jpg” and it is in the same directory. We need to use the image in grayscale mode, so give an extra parameter to load in grayscale mode.

# import the required modules
import cv2

# Read the image
img = cv2.imread('mic.jpg',0)

• We can reduce the noise in the image so that the round corners and edges are detected smoothly. We can use medianblur() method of cv2 to achieve that.

# Apply median blur
img = cv2.medianBlur(img,5)

• To detect the edges, we have various kind of thresholding functions available in openCV. But the most compatible one in this case is ADAPTIVE_THRESH_MEAN_C. Additionally, we can check other threshold functions too.

# Apply MEAN thresholding to get refined edges
image = cv2.adaptiveThreshold(img,255,cv2.ADAPTIVE_THRESH_MEAN_C,cv2.THRESH_BINARY,11,2)

Now we have an image having only edges. We can iterate over our image to print the bright and dark points and that will generate the dotted text for us. If the image is too large we can use cv2.resize() to shrink the image so that it can be used.

Implementation : 

Python3

# Python script to convert the given image
# into a dotted text using opencv

# import the required modules
import cv2

# Read the image
img = cv2.imread('mic.jpg', 0)

# Apply median blur
img = cv2.medianBlur(img, 5)

# Apply MEAN thresholding to get refined edges
image = cv2.adaptiveThreshold(
    img, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 11, 2)

# Convert the image into a compatible size
# We will use 60 pixels wide image so that text
# fits in the console

# Preserve the ratio
ratio = len(image)/len(image[0])
# Assign new width and calculate new height
new_width = 60
new_height = int(ratio*new_width)
# Resize the image
image = cv2.resize(image, (new_height, new_width))

# Iterate over the array and print the dark pixels
# or we can use any other symbol too.
for i in range(len(image)):
    for j in range(len(image[0])):
        print("o" if image[i, j] < 100 else ".", end="")
    print()

• Original Image : 

• Gray scale image after applying median blur :

• Image after applying Adaptive mean threshold :

• Final output printed on console :