Instance Segmentation Function

This module requires a neuron info dict from the instance segmentation step as well as the Allen labels directory (in native space), typically from the registration step.

This module perfoms the following:

1. Filters cells based on a specified minimum (and optional maximum) area argument

2. Computes cell count of segmented image and summarizes them per label

3. Outputs clarity_segmentation_features_ara_labels.csv (segmentation features summarized per ARA labels)

Main outputs

1. A csv file containing the segmentation features summarized per ARA labels located at <output_folder>/clarity_segmentation_features_ara_labels_<hemi>.csv

CLI

To look at the arguments that need to be provided to the function, invoke the help menu using:

$ miracl seg count_neurons -h

The following menu will be printed to the terminal:

usage: miracl count_neurons -l LBL --min-area MIN_AREA --neuron-info-dict
                        NEURON_INFO_DICT [-o OUTPUT] [--max-area MAX_AREA]
                        [--hemi {split,combined}] [--skip SKIP]
                        [-c CPU_LOAD] [-v] [-h]

1) Filters cells based on a specified minimum (and maximum) area argument 1)
Computes cell count of segmented image and summarizes them per label 2)
Outputs clarity_segmentation_features_ara_labels.csv (segmentation features
summarized per ARA labels)

optional arguments:
-o OUTPUT, --output OUTPUT
                        Output directory (default: .)
--max-area MAX_AREA   Maximum area of a neuron to be considered (default:
                        1000000)
--hemi {split,combined}
                        Hemisphere of the brain (split or combined) (default:
                        combined)
--skip SKIP           Number of slices to skip from start and end (default:
                        50) Note: this helps to skip the slices that are poor
                        quality due to the microscope
-c CPU_LOAD, --cpu-load CPU_LOAD
                        Percentage of CPU load to use (default: 0.3)
-v, --verbose         Prints out more information (default: False)
-h, --help            show this help message and exit

required arguments:
-l LBL, --lbl LBL     Allen labels directory (in native space) used to
                        summarize features; from registration step. Can also
                        be any other labels that you want to use to summarize
                        the features, such as cluster labels.
--min-area MIN_AREA   Minimum area of a neuron to be considered
--neuron-info-dict NEURON_INFO_DICT
                        Path to the neuron info dict from instance
                        segmentation (neuron_info_final.json)

Flag	Parameter	Type	Description	Default
-l, --lbl	LBL	str	Allen labels directory (in native space) used to summarize features; from registration step.	None (required)
--min-area	MIN_AREA	int	Minimum area of a neuron to be considered	None (required)
--neuron-info-dict	NEURON_INFO_DICT	str	Path to the neuron info dict from instance segmentation (neuron_info_final.json)	None (required)
-o, --output	OUTPUT	str	Output directory	./
--max-area	MAX_AREA	int	Maximum area of a neuron to be considered	1000000
--hemi	{split, combined}	str	Hemisphere of the brain (split or combined)	combined
--skip	SKIP	int	Number of slices to skip from start and end	50
-c, --cpu-load	CPU_LOAD	float	Percentage of CPU load to use	0.3
-v, --verbose	N/A	bool	Prints out more information	False

Note

The --skip argument is used to skip the slices that are poor quality due to the microscope.

Example usage:

$ miracl seg count_neurons \
    -l ../reag_final/allen_labels/ \
    --min-area 10 \
    --neuron-info-dict ./cc_patches/neuron_info_final.json \
    -o ./output_dir