Numpy Indexing with Arrays

Question

I have 2 arrays [nx1] that store xpixel (sample) and ypixel (line) coordinates, respectively. I have another array [nxn] storing an image. What I would like to do is create a third array which stores the pixel values from the image array at the given coordinates. I have this working with the following, but wonder if a built-in numpy function would be more efficient.

#Create an empty array to store the values from the image.
newarr = numpy.zeros(len(xsam))

#Iterate by index and pull the value from the image.  
#xsam and ylin are the line and sample numbers.

for x in range(len(newarr)):
    newarr[x] = image[ylin[x]][xsam[x]]

print newarr

A random generator determines the length of xsam and ylin along with the direction of travel through the image. It is therefore totally different with each iteration.

jorgeca · Accepted Answer · 2012-11-20 21:27:51Z

3

You can use advanced indexing:

In [1]: import numpy as np
In [2]: image = np.arange(16).reshape(4, 4)
In [3]: ylin = np.array([0, 3, 2, 2])
In [4]: xsam = np.array([2, 3, 0, 1])
In [5]: newarr = image[ylin, xsam]
In [6]: newarr
array([ 2, 15,  8,  9])

answered Nov 20, 2012 at 21:27

jorgeca

5,5223 gold badges26 silver badges36 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

jfs · Accepted Answer · 2012-11-20 21:32:07Z

3

If image is a numpy array and ylin, xsam are one dimensional:

newarr = image[ylin, xsam]

If ylin, xsam are two-dimensional with the second dimension being 1 e.g., ylin.shape == (n, 1) then convert them to one-dimensional form first:

newarr = image[ylin.reshape(-1), xsam.reshape(-1)]

answered Nov 20, 2012 at 21:32

jfs

417k210 gold badges1k silver badges1.7k bronze badges

1 Comment

jorgeca Over a year ago

You don't have to unravel ylin and xsam. If you don't do that, newarr will keep the same shape as ylin or xsam, which is quite useful (of course, the code from the OP returns a 1D newarr, but you could just .squeeze newarr at the end if you want that).

Collectives™ on Stack Overflow

Numpy Indexing with Arrays

2 Answers 2

Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related