Normalization (rescaling) of averages among cell

The second part of the workflow "normalizes" shapes of average BB volume images (per cell)
built during the previous part (see illustration below).

The reason is that per cell, average BB images are distorted by
slightly varying ExM factors and intracellular variability.
To compare and overlay protein of interest stains from different cells,
we'd need to adjust the shape and register reference "total protein stain" channel.

The workflow consists of four steps (see links below).
It is assumed that the input data/folder structure is the same as the first part's output.
There is a need for a manual correction after step 1, so it cannot
be run in a "fully automated" manner.

(part 2) Step 1: Segmenting basal body
(part 2) Step 2: Segmentation verification
(part 2) Step 3: Rescaling shapes
(part 2) Step 4: Registration

output folders

The output of this part is stored in a new separate "averaging" folder.
It has the following structure:
(averaged_data_folder corresponds to the Figure4/averaged_data_folder in the provided dataset)

averaged_data_folder
├── a1_measure
│   ├── BB_channel
│   │   └── ...
│   ├── BB_marks
│   │   └── ...
│   └── ...
├── a2_verification
│   ├── BB_marks
│   │   └── ...
│   ├── rescaleZ
│   │   └── ...
│   └── summary_XYZ_scales.csv
├── a3_rescaling
│   ├── not_rescaled
│   │   └── ...
│   └── rescaled
│       └── ...  
└── a4_averaged
    ├── not_rescaled_avrg
    │   └── ...
    └── rescaled_avrg
        └── ...