correcting unordered list intro

docs/intro.rst

@@ -17,22 +17,20 @@ This documentation intends to present all the processing chains required not onl
 
 The ``nipet`` (neuro-imaging PET) Python package contains all the routines needed for robust quantitative image reconstruction from raw data with added quality control of any processing segments, e.g., photon scatter or randoms corrections.  The raw data typically includes large list-mode (LM) data (usually taking up GBs), data for detector normalisation, and data for attenuation correction, i.e., the attenuation map, also called the |mu|-map.
 
-.. :math:`\mu`-map.
-
 The ``nimpa`` (neuro-image manipulation, processing and analysis) package contains all the routines for image input/output, image trimming and up-sampling for regional signal extraction (e.g. from a region of interest--ROI), image registration, and importantly, for partial volume correction (PVC). 
 
 The key aspect of ``nipet`` is the fast LM processing for efficient uncertainty estimation of statistic based on image or projection data.  Significant emphasis in placed on reconstruction for scanners with extended axial field of view (FOV), such as the latest PET/MR systems, e.g., the Siemens Biograph `mMR`_ or the GE `Signa`_ [*]_.
 
 The processing chains include:
 
-#. list-mode data processing;
-#. accurate attenuation coefficient map generation (with pseudo-CT for PET/MR);
-#. detector normalisation with dead time correction;
-#. exact forward and back projection between sinogram and image space;
-#. estimation of reduced-variance random events;
-#. high accuracy fully 3D estimation of scatter events;
-#. voxel-based partial volume correction;
-#. region- and voxel-level image analysis.
+  #. list-mode data processing;
+  #. accurate attenuation coefficient map generation (with pseudo-CT for PET/MR);
+  #. detector normalisation with dead time correction;
+  #. exact forward and back projection between sinogram and image space;
+  #. estimation of reduced-variance random events;
+  #. high accuracy fully 3D estimation of scatter events;
+  #. voxel-based partial volume correction;
+  #. region- and voxel-level image analysis.
 
 Due to its speed and additional functionalities, *NiftyPET* allows practical and efficient generation of multiple bootstrap realisations of raw and image datasets, being processed within arbitrarily complex reconstruction and analysis chains :cite:`Markiewicz2018b,Markiewicz2016`. Based on these datasets, distributions of any statistic can be formed indicating the uncertainty of any given parameter of interest, for example, the regional SUVr in amyloid brain imaging.
 
@@ -58,11 +56,14 @@ In stage **B**  the normalisation component data (relatively small file) is used
 Great emphasis was put on the quantitative image reconstruction and analysis in stages **D-H** (for more details see :cite:`Markiewicz2018b`):
 
   * forward and back projectors used for image reconstruction (stage **D**); the attenuation factors are generated with the forward projector.
+
   * fully 3D estimation of scatter events (stage **E**), with high resolution ray tracing in image and projection space; the  estimation is based on voxel-driven scatter model (VSM) and is coupled with image reconstruction, i.e., the scatter is updated every time a better image estimation of the radiotracer distribution is available.
+
   * voxel-wise partial volume correction using MRI brain parcellations (stage **F**), based on the iterative Yang method and given point spread function (PSF) of the whole imaging system (including the hardware and the reconstruction algorithm).
-  * kinetic analysis using dynamic multi-frame PET data (stage **G**)
-  * voxel-wise uncertainty estimation based on efficient generation of bootstrap LM data replicates (stage **H**). 
 
+  * kinetic analysis using dynamic multi-frame PET data (stage **G**).
+
+  * voxel-wise uncertainty estimation based on efficient generation of bootstrap LM data replicates (stage **H**). 
 
 
 .. rubric:: Footnotes