Version 2.2

This release fixes at least some of the bit-rot that has occurred since the last release.

Wheels of version 2.2 for Windows (64 bit) are available below.

Version 2.1

Major changes since version 2.0:

• 2x up-sampling was removed from the cubic spline.
• The localization fitting ROIs are now (approximately) round, the roundness is limited by pixelation.
• CMOS analysis was updated to handle pixel-to-pixel QE differences.

This is probably the last version that will explicitly support Python2.

Wheels of version 2.1 for Windows (64 bit) are available below.

Version 2.0

Major changes since version 1.1:

• The HDF5 format is used to store analysis results.
• Jupyter notebooks are available to describe / document the analysis.

Wheels of version 2.0 for Windows (64 bit) are available above.

Version 1.1.1

Major changes since version 1.0:

1. Image segmentation is done using the SNSMIL algorithm.
2. Localization fitting is done using the Levenberg–Marquardt algorithm.
3. The fit ROI of 3D-DAOSTORM / sCMOS is fixed (it used to auto-adjust up to a maximum size).

This is (probably) the last release that will use the Insight3 binary format. All versions after this point will use HDF5 to store analysis results with the option to convert the results to the Insight3 binary format.

Wheels of version 1.1.1 for Windows (64 bit) are available above.

Version 1.0

WARNING This version is not compatible with current numpy / Python.

This is the last version of the software with a fixed threshold for localization identification. All versions after this point will use a relative threshold. The localization finding and fitting programs will also require reasonably accurate estimates of the factors to convert camera ADUs to actual counts in photo-electrons.

Wheels of version 1.0 for Windows (64 bit) are available above.