Jitterbug is a piece of STEM image correction code, written in MatLab™, for the identification and correction of image distortions and drift. Distortion can be common in recorded STEM images and degrade signal-noise (SNR) or resolution performance. Compensating for these distortions can recover much of the lost performance and the code has so far demonstrated improvements of SNR and resolution of up to 42% and 10% respectively. Academic/non-commercial users can now Download Jitterbug for FREE. The current v3.6 release includes several new features. Now from a single STEM image the Jitterbug code can:
- Identify and compensate for horizontal and vertical ‘scan noise’,
- Measure sample or stage drift during acquisition and correct for it,
- Calculate the probe’s distortion frequency spectrum and identify the five strongest peaks,
- Quantify the resolution and signal-noise ratio of the raw and restored images.
Example ‘Before and After’ Results
Scanning transmission electron microscopy (STEM) can be affected by the effects of environmental instabilities because of the technique’s serial acquisition recording. Any time varying disturbance such as acoustic or seismic vibrations or electromagnetic fields can disturb the image recording process. The purpose of this image-processing code is to detect and correct for such image distortions.
The example to the left shows an enlargement from a section of a STEM HAADF image from  oriented Strontium Titanate (SrTiO3).
In the as-recorded image many atom columns appear sliced horizontally. In addition some atoms appear to have dissociated rows above or below their main form. In the restored image the image rows of each of the atom columns are brought back together. Additionally the image shear imparted from sample/stage drift has been corrected such that the lattice planes once more appear perpendicular as expected. In the example (left) the Jitterbug restoration code improved SNR and resolution by 30% and 13% respectively
In the Fourier Transform (FT) of the as-recorded image there are two strong bands running vertically that arise from the scan noise as well as streaking of the Fourier spots. In the restored FT these band are greatly reduced as well as the spot streaking. The drift compensation is also evident in the FT where, after reconstruction, FT spots are correctly at ninety degrees from one another.
A full-length manuscript describing this work was published in the journal Microscopy & Microanalysis and can be downloaded here.
“…steps have been taken to ameliorate these complications through the use of correcting algorithms […]. These image processing techniques are quite powerful and ensure maximal data veracity before the analysis is initiated” – Alex Belianinov et al. Nature Communications, 6 7801 DOI:10.1038/ncomms8801
“Strategies to average, or remove, distortions in HAADF STEM images are now available, allowing strain analysis to be employed with confidence” – Jonathan Peters et al. Ultramicroscopy 157, p.91–97 DOI:10.1016/j.ultramic.2015.05.020
|Feature||MatLab Version||Digital Micrograph Plug-in|
(units of pixels only)
(units of nanometres)
|Instability frequency analysis|
|Bright-field / ABF support|
|Interactive user interface|
|Large image support|
(not recommended above 512x512px)
(tested up to 2048x2048px)
|*.dm4 file support|
academic use only
|Fully licensed for academic
or commercial use
|Availability||Download below||Contact: email@example.com|