Napari-Correct-Drift’s doc

Fiji’s Correct-3D-drift macro for Napari

In time-lapse imaging, a motorized microscope repeatedly captures images at specified positions for long periods. However, microscope stages exhibit drift, causing the sample to move, or to appear moving.

The reasons are complex and manifold, but it typically results from error propagation in odometry, thermal expansion, and mechanical movement. Drift poses problems for subsequent image analysis and needs to be corrected.

With Napari-correct-drift provides an extensible solution with similar functionality as Fiji’s Correct-3D-drift macro. It offers efficient cross-correlation using Fourier phase correlation method, improved key frame selection, and outlier handling. In Napari users can provide a regions-of-interest (ROI) to effectively stabilize objects in up-to 3D-multi-channel images. Additionally, estimated drifts can be exported, imported, or edited before applying the correction.

When to use this plugin

  1. Your time-series images or volumes exhibit translational drift, i. e. rigid movement, without rotation.

  2. Reference channel with fixed object (e. g. fiducial) visualizing the drift

  3. Stabilizing objects of interest by using a ROIs

Without Napari viewer

Napari-Correct-Drift can also be used without starting the Napari viewer.

from napari_correct_drift import CorrectDrift

# multi-channel 2D-movie
cd = CorrectDrift(img_in, "tcyx")

# estimate drift table
drifts = cd.estimate_drift(t0=0, channel=0)

# correct drift
img_cor = cd.apply_drifts(drifts)

With Napari viewer (using ROI)

Stabilizing an growing root-tip using an ROI.

Test data

Napari-correct-drift contains synthetic sample data. To test it on real data download an example Arabidopsis growing root tip

Issues and contributing

If you have any problems or question running the plugin, please open an issue

Contents:

Indices and tables