DLASLBuildMask
==============

**Purpose.** Build a DL-ASL mask by combining tissue segmentations (C1, C2, C3), then **resample to 64×64×24** for the CNN-based denoiser.

Inputs & discovery
------------------

- Looks in ``derivatives/<sub>/perf`` for files containing ``c1``, ``c2``, ``c3`` (``*.nii`` or ``*.nii.gz``).
- Uses the first subject/session from ``data_descrip["Images"]``.

Outputs
-------

- Binary mask at 64×64×24: default ``derivatives/<sub>/perf/dlasl_mask.nii.gz``

Configuration
-------------

=================  =========================  ======================================
Key                Type / Default             Meaning
=================  =========================  ======================================
``out_mask``       str / ``dlasl_mask.nii.gz``  Output filename (absolute or relative to ``perf``)
=================  =========================  ======================================

YAML example
------------

.. code-block:: yaml

    modules:
      - name: DLASLBuildMask
        params:
          out_mask: dlasl_mask.nii.gz