Resampling does not exactly match fmriprep output

I just ran fmriprep with --output-spaces T1w and I was wondering how the sub-001*space-T1w_desc-aparcaseg_dseg.nii.gz in the func folder was generated?

Is it simply a resampling of the sub-001_desc-aparcaseg_dseg in the anat folder?

I tried to resample sub-001_desc-aparcaseg_dseg.nii.gz with nilearn resample_to_img tool but the output, even though close, is not identical to sub-001*space-T1w_desc-aparcaseg_dseg.nii.gz.