Data Scientist Guide¶

CloudScope is organized around a small scientific data model that supports both GUI workflows and Python scripting.

The main scientific objects are:

AcqImage: one acquisition-backed image dataset
AcqImageList: a collection of AcqImage objects
ROI: the region where analysis is performed
Analysis: a quantitative workflow applied to an image, channel, and ROI

Current quantitative analysis workflows are designed for line scan kymographs and include:

blood flow velocity analysis from line scan kymographs using a Radon-transform-based method
vessel diameter analysis from line scan kymographs

Why CloudScope?¶

CloudScope exists to keep scientific image loading, visualization, metadata, ROIs, and analysis connected through one shared backend.

The same acqstore code is used by:

the desktop GUI
the browser GUI
Python scripts
Jupyter notebooks
unit tests

This makes it possible to reproduce GUI analysis results programmatically using the same underlying APIs.

Core concepts¶

AcqImage¶

An AcqImage represents one acquisition-backed image dataset.

An AcqImage combines:

image pixels
image header metadata
experimental metadata
ROIs
analysis summaries and results

See the AcqImage API.

AcqImageList¶

An AcqImageList manages a collection of AcqImage objects.

Typical uses include:

loading a folder of acquisition files
managing multiple images in one workflow
running batch analysis
building tables from file-level metadata

See the AcqImageList API.

ROIs¶

ROIs define the region of image data used for analysis. Analysis results are associated with an image, channel, and ROI.

Analysis¶

Analysis modules take detection parameters as input and produce summary values and optional tabular outputs. Detection parameters define scientific behavior; execution options control runtime behavior such as progress, cancellation, and parallel processing.

Line scan kymographs¶

The current quantitative analysis workflows operate on line scan kymographs rather than general two-dimensional image fields.

A line scan kymograph represents repeated sampling along a spatial line over time. CloudScope uses this structure for blood flow velocity and vessel diameter measurements.

Current analysis workflows¶

Blood flow velocity¶

Velocity analysis estimates blood flow velocity from line scan kymographs using a Radon-transform-based method.

Vessel diameter¶

Diameter analysis estimates vessel diameter from line scan kymographs using intensity-profile-based measurements.

Saved analysis files¶

For each acquisition file, CloudScope stores analysis state and results as sidecar files next to the source file.

For a source file named my_file.tif, CloudScope may save:

my_file.tif
my_file.tif.json
my_file.tif.radon_velocity.csv
my_file.tif.diameter.csv

The JSON sidecar stores:

accepted/rejected state
experimental metadata
image header metadata
ROIs
detection parameters used for each analysis
analysis summaries

The CSV files store tabular analysis outputs for analysis types that provide CSV export.

When CloudScope reloads an AcqImage, the JSON sidecar is used to restore relevant metadata, ROIs, and saved analysis state.

Metadata¶

CloudScope separates image header metadata from user-editable experimental metadata.

Image header metadata is read from the source file when available. Experimental metadata is edited by the user and saved with the AcqImage sidecar.

See AcqImage Metadata.

Current limitations¶

CloudScope can load and visualize supported scientific image formats, but the currently implemented quantitative analysis workflows are designed for line scan kymographs.

Traditional two-dimensional segmentation, tracking, and image-analysis workflows are not yet implemented as dedicated CloudScope analysis modules.