Using a Spacing class for changing spacing of a CT only does changes in the shape of the image(as expected). But cannot change the metadata information for spacing of output data.

[one2clouds]

Hello,
the following is my code

self.input_img_name = './001.mha'

required_spacing = (1.0, 1.0, 1.0)

input_arr = LoadImage(image_only=True)(self.input_img_name)

spacer = Compose([ LoadImaged(keys="image"), Spacingd(keys="image", pixdim=required_spacing, mode="bilinear")])

output_arr = spacer({"image":self.input_img_name})

print(input_arr.shape) # torch.Size([512, 512, 401])
print(output_arr["image"].shape) #torch.Size([512, 400, 314])
print(output_arr["image_meta_dict"]["spacing"]) # [0.78125    0.78125    0.80000001]

From the last and the second last line.
We can see that the output array has changed but, the meta data of the output for the spacing remains the same.
[It should be (1.0,1.0,1.0) but we got (0.78, 0.78, 0.8)]
Its the same case for the orientation, origin, spacing.
How can we get transformed orientation, origin and spacing from output meta data?

[KumoLiu]

Hi @one2clouds, may I ask about the MONAI version you used?