Get matrix image of numpy array values - Grid with pixel values inside (not colors)

Question

I searched online and couldn't find anything about this that does what I want.

I would like to save a numpy array as an image but instead of having a colorful image, I want a black and white representation of the pixel values in their corresponding grid location.

For example:

import numpy as np
x = np.array([[1,2],[3,4]])
print(x)
# [[1 2]
#  [3 4]]

I would like to save this as an image (.PNG) that looks like the following:

My current code creates a grid and places the numbers inside but it is very difficult to adjust everything to make it presentable in a research paper.

So rather than posting my overly complex code, I was wondering if there is a built in function to handle this in a few lines of code.

Looks like table would be a good place to start. I haven't used it enough to give a good answer. — Warren Weckesser
– Warren Weckesser, Commented Nov 27, 2019 at 22:22
I would use LaTeX for a table you're trying to put into a paper, as opposed to a raster image. — alkasm
– alkasm, Commented Dec 5, 2019 at 7:02
@alkasm the thing is my actual matrices are much larger (14x14 or 28x28) and I need to make many images for them, so manually doing so in LaTex seems like a very inefficient method. — lbragile
– lbragile, Commented Dec 5, 2019 at 7:03
On the contrary, rasterizing to images sounds like the inefficient method. Tables of data are not images, and so should not be treated as such. You can auto generate LaTeX, even manually, extremely easily. — alkasm
– alkasm, Commented Dec 5, 2019 at 23:33

Solvalou · Accepted Answer · 2019-12-06 05:41:25Z

I would use LaTeX to generate the tables, since they look fancy and you can either generate an image or directly put them in your document. I used the following code to achieve this:

#!/usr/bin/env
import numpy as np
import os

x = np.array([[1,2],[3,4]])

def generateLatexTable(x):
    start = [r'\documentclass[preview]{standalone}', r'\begin{document}', r'\begin{tabular}{%s}' % ('{1}{0}{1}'.format('|'.join(['r'] * x.shape[1]), '|')), r'\hline']
    tab = [' & '.join(['%d' % val for val in row]) + r' \\ \hline' for row in x]
    end = [r'\end{tabular}', r'\end{document}']
    text = '\n'.join(start + tab + end)

    return text

with open('table.tex', 'w') as f:
    f.write(generateLatexTable(x))
os.system("pdflatex table.tex")

Here, the document class preview is used which returns an image resized to the content of the document, i.e. just the table. Only a tabular environment is used to present the data. There are horizontal and vertical bars between the cells, but it is very easy to change this. In the variable tab the data is processed for each row and converted into a string. Note that you have to specify the output format at this position. I set it to %d so everything is converted to integers.

If you want to use the table directly in a latex source, you have to remove documentclass and \begin{document} as well as \end{document} in the variables of start and end. Finally, everything is put together in a latex-source which is then stored to disk as table.tex. If you just want the image in the end, the resulting file is compiled to table.pdf.

Here is what the output looks like. But like I said, it is very easy to change the looks since it is LaTeX :)

Here is another example with a large matrix (14 x 14), filled with random numbers ranging from 0 to 100:

Ashwin Geet D'Sa · Accepted Answer · 2019-12-11 10:20:24Z

8

You can use the table function of matplot to plot the simple table. Furthermore, you can save the plot as PNG. Below is the simple code for your requirements:

import numpy as np
import matplotlib.pyplot as plt

x = np.array([[1,2],[3,4]])

plt.figure()
plt.table(cellText=x,cellLoc='center',loc='center')
plt.axis('off')
plt.savefig('table.png')

Size of the plot or image can be adjusted by changing figsize parameters in the line : plt.figure(figsize=(x,y))

For better appearance, it can be modified as below:

import numpy as np
import matplotlib.pyplot as plt

x = np.array([[1,2],[3,4]])

fig = plt.figure(figsize=(2,2))
plt.axis('off')
plt.axis('tight')
plt.table(cellText=x,cellLoc='center',loc='center')
#plt.subplots_adjust(hspace=0.5)
fig.tight_layout()
plt.savefig('table.png')

edited Dec 11, 2019 at 10:20

answered Dec 5, 2019 at 9:59

Ashwin Geet D'Sa

7,4583 gold badges40 silver badges66 bronze badges

1 Comment

Daniel F Over a year ago

I'd almost do something like fig = plt.figure(figsize = x.shape)

Ivan · Accepted Answer · 2019-12-05 09:03:17Z

4

May be this will help:

from matplotlib import pyplot as plt
import numpy as np

w  = 10
h  = 10 
img = np.random.randint(255, size=(w, h))
plt.figure(figsize=(5,8))
plt.imshow(img, interpolation='nearest')
plt.axis('off')

cellTextimg =  []
for j in range(0,h):
    cellTextimg.append(img[j,:])

the_table = plt.table(cellText= cellTextimg, loc='bottom')

answered Dec 5, 2019 at 9:03

Ivan

1,5512 gold badges20 silver badges42 bronze badges

Collectives™ on Stack Overflow

Get matrix image of numpy array values - Grid with pixel values inside (not colors)

3 Answers 3

Comments

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related