Plot colour image from a numpy array that has 3 channels

Question

in my Jupyter notebook I am trying to display an image that I am iterating on through Keras. The code I am using is as below

def plotImages(path, num):
 batchGenerator = file_utils.fileBatchGenerator(path+"train/", num)
 imgs,labels = next(batchGenerator)
 fig = plt.figure(figsize=(224, 224))
 plt.gray()
 for i in range(num):
    sub = fig.add_subplot(num, 1, i + 1)
    sub.imshow(imgs[i,0], interpolation='nearest')

But this only plots single channel, so my image is grayscale. How do I use the 3 channels to output a colour image plot. ?

So just do sub.imshow(imgs) to use all channels. How do you expect it to show RGB if you don't provide all channels? — Suever
– Suever, Commented Jan 25, 2017 at 16:38

Suever · Accepted Answer · 2017-01-25 17:55:06Z

11

If you want to display an RGB image, you have to supply all three channels. Based on your code, you are instead displaying just the first channel so matplotlib has no information to display it as RGB. Instead it will map the values to the gray colormap since you've called plt.gray()

Instead, you'll want to pass all channels of the RGB image to imshow and then the true color display is used and the colormap of the figure is disregarded

sub.imshow(imgs, interpolation='nearest')

Update

Since imgs is actually 2 x 3 x 224 x 224, you'll want to index into imgs and permute the dimensions to be 224 x 224 x 3 prior to displaying the image

im2display = imgs[1].transpose((1,2,0))
sub.imshow(im2display, interpolation='nearest')

edited Jan 25, 2017 at 17:55

answered Jan 25, 2017 at 16:43

Suever

65.6k14 gold badges91 silver badges104 bronze badges

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

4 Comments

Abhik Over a year ago

I am getting an error , 'TypeError: Invalid dimensions for image data'

Suever Over a year ago

@Abhik You said your image was 244 x 244 x 3. What does imgs.shape show?

Abhik Over a year ago

(2, 3, 224, 224) - 2 images of 3 *224*224 . Hence I am doing sub.imshow(imgs[i], interpolation='nearest') , where i is the index of the image .

sh37211 Over a year ago

TypeError: transpose() received an invalid combination of arguments - got (tuple), but expected one of:  * (int dim0, int dim1)  * (name dim0, name dim1)

HRn · Accepted Answer · 2021-02-28 21:30:59Z

0

It will work for you: Try putting the channels last by permitting,

image.permute(2 , 3 , 1 , 0)

and then remove the image index by np.squeeze():

plt.imshow((np.squeeze(image.permute(2 , 3 , 1 , 0))))

answered Feb 28, 2021 at 21:30

HRn

1

Collectives™ on Stack Overflow

Plot colour image from a numpy array that has 3 channels

2 Answers 2

4 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related