Fixed
* Header: When using a template to create DicomHeaderDict, use deepcopy to get a unique
DicomHeaderDict for the new instance.
Added
* Viewer: Save window center/width to Series object when returning.
* Series.fuse_mask(): New parameter `blend` which determines whether
the self image is blended. Default is False, such that the self image is
not blended, only the mask is blended.