Quick Start

Introduces the basic usage of MitoVis through a scenario of
1) loading a pre-generated dataset into MitoVis,
2) manually modifying the neuronal structure and mitochondria object, and
3) analyzing it.

(Deep learning environment setting and deep learning related functions (initial dataset generation, interactive model fine-tuning, projection...) will be introduced later)

1. Requirements

OS: Windows 10, 11 tested
Microsoft Visual C++ 2017
(If this is not installed, run (vc_redist.x64.exe) to install it)

2. Start

To get started, download the MitoVis_QuickStart version from the Download page.

Unzip the downloaded zip file, and open the MitoVis.exe.

Then, we can see the initial layout of MitoVis.

3. Load Dataset

1) Press the plus button to add a new group

2) Press the plus button on the added group to load dataset

3) Select CON_60X_1.nd2.MitoVis file in the con60x folder

The pre-generated dataset includes neuron structure image, mitochondria image, neuronal structure segmentation label, and mitochondria objects label.

We can
1) observe the images and labels on the image viewer,
2) adjust imaging parameters on the control panel,
3) observe and filter the features of mitochondria objects on the analysis panel,
4) filter the mitochondria objects to generate an analysis subset with feature-based filtering on the analysis panel and image-based filtering on the image viewer,
5) summarize result of the subset on the morphology comparison panel.

We can also adjust the layout by dragging the handles

4. Data Exploration

Image Viewer

Zoom in/out: mouse right button drag up/down
Moving: mouse left button drag
Moving (when some function is activated): space key + mouse left button drag
Measuring: mouse middle button drag
Object focusing: mouse hover on a mitochondria object for focusing the related feature on the plots

Parallel Coordinate Plot

Range filtering: mouse right button drag on each axis
Threshold filtering: move the arrows on the end of each axis
Object focusing: mouse hover/click the target line

2D Projection Plot

Axis setting: change the feature of drop boxes (X-Axis, Y-Axis)
(for PCA or UMAP, python [version: >3.6] and some libraries [scikit-image, and umap-learn] are required)
Threshold filtering: move the arrows on the side of plot
Object focusing: mouse hover/click the target point

Control Panel (Neuron Channel / Mito Channel / Label Channel)

Image visualization parameters for each channel can be adjusted by modifying sliders and check boxes

Control Panel (ROI Selection)

Activate the Add/ Erase button and brush desired region for image-based mitochondria object filtering

Morphology Comparison Panel

Summarizes average features of current analysis subset (activated mitochondria objects)

Summary generation: press the plus button on the left side
Saving summaries: press the disk button on the left side
Comparison bar chart generation: set the number of FV1 and FV2 (the value of bar chart is calculated as [(1 - FV2/FV1) * 100%])

5. Neuronal Structure Correction

6. Mitochondria Object Correction

Google Sites

Report abuse