Advanced electron holography: a new algorithm for image processing and a standardized quality test for the FEG electron microscope

Völkl, E.; Allard, L. F.; Datye, Abhaya K.; Frost, Bernhard G.

doi:10.1016/0304-3991(94)00182-m

Cited by 20 publications

(9 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…SNR is the signal-to-noise ratio in the hologram, N el is the number of electrons collected per pixel and I max and I min are the maximum and minimum intensities of the interference fringes, respectively (Völkl et al, 1995). Figure 3c shows measurements of fringe visibility plotted as a function of biprism voltage for an FEI Titan TEM operated at different magnifications.…”

Section: Interference Fringe Visibilitymentioning

confidence: 99%

Electron Holography of Magnetic Materials

Kasama¹,

Dunin‐Borkowski²,

Beleggia³

2011

Holography - Different Fields of Application

View full text Add to dashboard Cite

Section: Interference Fringe Visibilitymentioning

confidence: 99%

Electron Holography of Magnetic Materials

Kasama¹,

Dunin‐Borkowski²,

Beleggia³

2011

Holography - Different Fields of Application

View full text Add to dashboard Cite

“…This question was answered essentially in [3] and the reader inclined to review the mathematics is recommended to read these articles. However, there it was only shown how to shift a Fourier transform by an arbitrary amount.…”

Section: Shifting Images With Sub-pixel Accuracymentioning

confidence: 99%

“…The data acquisition "piggybacks" onto the live data acquisition from DM and provides a 3D-data set with n slices, with each of the n slices being a full image. In a second step, the slices in the 3Ddata set are aligned laterally with respect to each other (at sub-pixel accuracy) [3]. Last, the data are averaged yielding a single, noise improved image.…”

mentioning

confidence: 99%

Improving Image Quality and Reducing Drift Problems via Automated Data Acquisition and Averaging in Cs-corrected TEM

et al. 2008

View full text Add to dashboard Cite

IntroductionImage acquisition with a CCD camera is a single-press-button activity: after selecting exposure time and adjusting illumination, a button is pressed and the acquired image is perceived as the final, unmodified proof of what was seen in the microscope. Thus it is generally assumed that the image processing steps of e.g., "dark-current correction" and "gain normalization" do not alter the information content of the image, but rather eliminate unwanted artifacts. Image quality therefore is, among a long list of other parameters, defined by the dynamic range of the CCD camera as well as the maximum allowable exposure time depending on sample drift (ignoring sample damage). Despite the fact that most microscopists are satisfied with present, standard image quality we found that it is a relatively easy to improve on existing routines in at least two aspects [1]:1. Suppression of lateral image drift during acquisition by using significantly shorter exposure times with a plurality of exposures (3D-data set). 2. Improvement in the Signal/Noise ratio by averaging over a given data set by exceeding the dynamic range of the camera. BackgroundImages recorded in an electron microscope are usually band-width limited. This can be verified easily by taking a Fourier transform (FFT) of a given image and checking that no structures contribute visibly at the outer edges of the Fourier transform (also known as the Nyquist limit [2]). Otherwise, the image could contain artifacts and should be discarded if in doubt. For the following, it will be assumed that the acquired images are of reasonable quality and thus inherently bandwidth limited at the Nyquist limit. As a reminder, a digital image consists of M×N, usually quadratic, pixels with distance d between consecutive pixels in x-and y-direction. Most importantly: each pixel represents a point and not an image area! This little detail sometimes is conveniently overlooked (mostly without causing any damage), but should be remembered when the relation between the image intensity I(x,y) as a function of x-and y-coordinates, versus the pixel values P m,n that represent the intensity function at selected points with coordinates (m,n) needs to be considered.

show abstract

“…There are two basic varieties of apodization functions: hard filters, such as the Hann window [1], where the magnitude is forced to zero outside the radius of the function, and soft filters, such as a Lorentz window [3]. Typically the radius of a soft apodization function must be considerably lower than a hard apodization to avoid an unacceptable amount of the autocorrelation and conjugate sideband being contained in the reconstruction and thereby introducing periodic artifacts.…”

mentioning

confidence: 99%