CODEWITHSUNDEEP
pythontensorflowkerasimagingmedical-imaging
psj
psj

Reputation: 376

dimensions of image read using tfio.image.decode_dicom_image

I read a .dcm image from my own dataset using tfio.image.decode_dicom_image following this link. The shape of the read image is shown as (None,None,None,None). While displaying the image using plt.imshow, the output of np.squeeze is zero-dimensional and hence giving error in 2-d image display. Could someone please help me understand what is the problem?

Upvotes: 0

Views: 198

Answers (2)

Mehmet Burak Sayıcı
Mehmet Burak Sayıcı

Reputation: 467

I think what you've missed is reading the file first. I was also thinking that tfio.image.decode_dicom_image will eliminate the need for the tf.io.read_file, but no.

     import tensorflow_io as tfio
     
     image_bytes = tf.io.read_file('dicom_00000001_000.dcm')
     
     image = tfio.image.decode_dicom_image(image_bytes, dtype=tf.uint16)

Reference : https://www.tensorflow.org/io/tutorials/dicom?hl=tr

Upvotes: 0

Kukesh
Kukesh

Reputation: 510

Try pydicom to read .dcm data

from pydicom import dcmread
import matplotlib.pyplot as plt     

read= dcmread('000000.dcm')
img=read.pixel_array
plt.imshow(img,cmap = 'gray')

or sometime the directiory is wrong

Upvotes: 0

Related Questions