SaveImaged not saving the correct metadata of the image

[ArpanGyawali]

I have used monai-deploy-app-sdk but I am having problem with the monai transforms. So I am here asking for the help.
Here I have created a skull strip app that takes nifty as an input. So i created a niftyloader_operator that loads the image data and the metadata from the file and is given to SkullStrip_operator which performs pretransforms and posttransforms.

The code for NiftyLoader_operator is something like this
import numpy as np

import monai.deploy.core as md
from monai.deploy.core import DataPath, ExecutionContext, InputContext, IOType, Operator, OutputContext, Image

from monai.utils import MetaKeys, SpaceKeys

from monai.deploy.utils.importutil import optional_import

nib, _ = optional_import("nibabel")


@md.input("image_path", DataPath, IOType.DISK)
@md.output("image", Image, IOType.IN_MEMORY)
@md.env(pip_packages=["SimpleITK>=2.0.2", "nibabel"])
class NiftiDataLoader(Operator):


    def compute(self, op_input: InputContext, op_output: OutputContext, context: ExecutionContext):
        input_path = op_input.get().path
        image = self.convert_and_save(input_path)
        op_output.set(image)



    def convert_and_save(self, nii_path):

        img = nib.load(nii_path)
        header = self._get_meta_dict(img)
        header[MetaKeys.AFFINE] = self._get_affine(img)
        data = self._get_array_data(img)
        return Image(data, header)
    
    def _get_meta_dict(self, img) -> dict:

        # swap to little endian as PyTorch doesn't support big endian
        header2 = dict()
        try:
            header1 = img.header.as_byteswapped("<")
        except ValueError:
            header1 = img.header
        header1 = dict(header1)
        header2["row_pixel_spacing"] = header1["pixdim"][1].item()
        header2["col_pixel_spacing"] = header1["pixdim"][2].item()
        header2["depth_pixel_spacing"] = header1["pixdim"][3].item()
        header1.update(header2)
        return header1
    
    def _get_affine(self, img):

        return np.array(img.affine, copy=True)
    
    
    def _get_array_data(self, img):

        data = np.asanyarray(img.dataobj, order="C")
        return data.transpose(2, 1, 0)

And the Pre transforms performed in SkullStrip_operator are

Compose(
                [
                    LoadImaged(keys=my_key, reader=img_reader),
                    SaveImaged(
                        keys=my_key,
                        meta_keys="image_meta_dict",
                        writer="NibabelWriter",
                        resample=False,
                        output_dir=out_dir,
                        output_postfix="temp",
                    ),
                    EnsureTyped(keys=my_key),
                    EnsureChannelFirstd(keys=my_key),
                    Orientationd(keys=my_key, axcodes="RAS"),
                    Spacingd(keys=my_key, pixdim=[1.0, 1.0, 1.0], mode="bilinear"),
                    ScaleIntensityRanged(keys=my_key, a_min=-200, a_max=1000, b_min=0.0, b_max=1.0, clip=True),
                ]
             )

Here LoadImaged is a custom InMemImageReader

Issue
Here, when i run the code using monai-deploy exec I get this sort of output in the terminal

Going to initiate execution of operator NiftiDataLoader
Executing operator NiftiDataLoader (Process ID: 133827, Operator ID: c97697f3-045e-4dc2-b0ab-e1fd1f5f6a7f)
Done performing execution of operator NiftiDataLoader

Going to initiate execution of operator SkullStripOperator
Executing operator SkullStripOperator (Process ID: 133827, Operator ID: 80820a44-f518-4def-81c0-343919c1672c)
Converted Image object metadata:
sizeof_hdr: 348, type <class 'numpy.ndarray'>
data_type: b'', type <class 'numpy.ndarray'>
db_name: b'', type <class 'numpy.ndarray'>
extents: 0, type <class 'numpy.ndarray'>
session_error: 0, type <class 'numpy.ndarray'>
regular: b'r', type <class 'numpy.ndarray'>
dim_info: 0, type <class 'numpy.ndarray'>
dim: [  3 512 512 705   1   0   0   0], type <class 'numpy.ndarray'>
intent_p1: 0.0, type <class 'numpy.ndarray'>
intent_p2: 0.0, type <class 'numpy.ndarray'>
intent_p3: 0.0, type <class 'numpy.ndarray'>
intent_code: 0, type <class 'numpy.ndarray'>
datatype: 4, type <class 'numpy.ndarray'>
bitpix: 16, type <class 'numpy.ndarray'>
slice_start: 0, type <class 'numpy.ndarray'>
pixdim: [-1.         0.5449219  0.5449219  0.5        0.         0.
  0.         0.       ], type <class 'numpy.ndarray'>
vox_offset: 0.0, type <class 'numpy.ndarray'>
scl_slope: nan, type <class 'numpy.ndarray'>
scl_inter: nan, type <class 'numpy.ndarray'>
slice_end: 0, type <class 'numpy.ndarray'>
slice_code: 0, type <class 'numpy.ndarray'>
xyzt_units: 10, type <class 'numpy.ndarray'>
cal_max: 0.0, type <class 'numpy.ndarray'>
cal_min: 0.0, type <class 'numpy.ndarray'>
slice_duration: 0.0, type <class 'numpy.ndarray'>
toffset: 0.0, type <class 'numpy.ndarray'>
glmax: 0, type <class 'numpy.ndarray'>
glmin: 0, type <class 'numpy.ndarray'>
descrip: b'TE=3e+02;Time=214037.142', type <class 'numpy.ndarray'>
aux_file: b'Sn_Arterial_Phase', type <class 'numpy.ndarray'>
qform_code: 1, type <class 'numpy.ndarray'>
sform_code: 1, type <class 'numpy.ndarray'>
quatern_b: 0.0, type <class 'numpy.ndarray'>
quatern_c: 1.0, type <class 'numpy.ndarray'>
quatern_d: 0.0, type <class 'numpy.ndarray'>
qoffset_x: 129.2275390625, type <class 'numpy.ndarray'>
qoffset_y: 81.7724609375, type <class 'numpy.ndarray'>
qoffset_z: -419.5, type <class 'numpy.ndarray'>
srow_x: [ -0.5449219   0.         -0.        129.22754  ], type <class 'numpy.ndarray'>
srow_y: [-0.         0.5449219 -0.        81.77246  ], type <class 'numpy.ndarray'>
srow_z: [   0.     0.     0.5 -419.5], type <class 'numpy.ndarray'>
intent_name: b'', type <class 'numpy.ndarray'>
magic: b'n+1', type <class 'numpy.ndarray'>
row_pixel_spacing: 0.544921875, type <class 'float'>
col_pixel_spacing: 0.544921875, type <class 'float'>
depth_pixel_spacing: 0.5, type <class 'float'>
affine: [[  -0.54492188    0.           -0.          129.22753906]
 [  -0.            0.54492188   -0.           81.77246094]
 [   0.            0.            0.5        -419.5       ]
 [   0.            0.            0.            1.        ]], type <class 'numpy.ndarray'>
2023-11-01 11:33:07,592 INFO image_writer.py:194 - writing: /data/gpfs/projects/punim0406/Projects/skull_strip_deploy/N023_test/for_nifty/nifty_temp/Img_in_context/Img_in_context_temp.nii.gz

Where the metadata are correctly shown, but the SaveImaged that is immidiately after LoadImaged provides you a new nifty data with completely different metadata than the loaded image itself.
It has origin:(0,0,0) and spacing(1,1,1).

Why is this happening. Why is the SaveImaged not saving the correct metadata?
I really need the help. I tried everything I could. Thank you in advance.

[KumoLiu]

Hi @ArpanGyawali, I noticed that there is an Orientationd and Spacingd after SaveImaged.

[ArpanGyawali]

Yes, its because i wanted to save the image that was loaded and then apply other transformation.

[KumoLiu]

Hi @ArpanGyawali, I can't reproduce the issue. The meta information is also saved within SaveImaged. How did you check the meta information?

import torch
import monai.transforms as mt
from monai.data import MetaTensor

affine = torch.eye(4)
affine[0, 0] = -12
affine[1, 1] = 2

test_data = MetaTensor(torch.rand(1, 3, 16, 16), affine=affine)
mt.SaveImaged(
        keys="image",
        meta_keys="image_meta_dict",
        writer="NibabelWriter",
        resample=False,
        output_dir="/workspace/",
        output_postfix="temp",
    )({"image": test_data})
print(mt.LoadImage()("/workspace/0/0_temp.nii.gz").affine)