Stripper

STRIPPER: faST Robust fIlament Picking ProcEduRe

Helps you to strip your filaments out of the micrograph automatically :-)

Overview

Stripper is the filament particle picking procedure used in our lab.

Installation

Activate the “biomedgroup” update site and add a new update site:

   http://sites.imagej.net/Mpi-do-stripper/

Get your data into ImageJ

Load the raw data

  1. File → Import → Image Sequence
  2. Select the folder of your images
  3. Type the file ending into “File name contains:” (e.g .mrc, .tif)
  4. Check “Use virtual stack”
  5. Press OK Now you have read in your data as an image stack.

Reduce the size

For the sake of memory consumption and speed your should reduce the size / image depth of your image data.

  1. Image → Adjust → Size
  2. Set the width to 1024. The height is calculated automatically.
  3. Press OK.
  4. Image → Type → 8 Bit Wait until the processing is finished.

Save the file for later use

File → Save As → Tiff…

Parameters

Detection

As there is no easy way to guess the lower and upper threshold, it is recommended to use the detection assistant (File → Detection assistant).

Filament width:

The width of your filaments in pixel.

Mask width (Advanced option):

As the response along the filament randomly fluctuates and sometimes disappear, it is important to average along the filament.

Lower threshold:

Filter responses lower than that value will be considered as background and filament tracing stops here.Higher values lead to more segmented lines, as low contrast regions of a filament will be considered as background the line is splitted at this position.

Upper threshold:

Filter responens higher than that will be considered as valid filament starting point.With higher values the contrast of the filaments have to higher be to be picked.

Equalize:

After enhancement of the filaments, all images are adjusted that filamants have a more similar absolute response. Especially when the image contains high contrast contaminations, this option has a positive effect. By defaut this option is activated.

Filtering

Min number of boxes:

The minimum number of boxes that have to be placed per line.


Response filter

Sigma min. response:

The mean response M and the standard deviation S are estimated over all lines per image. If a line has a signficant number of line points below the threshold T = M-F*S it is considered as false-positiv (background) and is removed, whereas F is value to specifiy.By default, this filter is deactivated (set to 0). A resonable value lies typically between 2 - 3.

Sigma max. response:

The mean response M and the standard deviation S are estimated over all lines per image. If a line has a signficant number of line points above the threshold T = M+F*S it is considered as false-positiv (ice,overlapping filament) and is removed, whereas F is value to specifiy.By default, this filter is deactivated (set to 0). A resonable value lies typically between 2 - 3.

Sensitivity:

Threshold value for minimum straightness. Line segments (see window size) with a straightness below that threshold will be removed and the line is splitted. Higher values means that the filament have to be more straight, where 1 means a perfect straight line.


Straightness filter

Min. straightness:

Threshold value for minimum straightness. Line segments (see window size) with a straightness below that threshold will be removed and the line is splitted. Higher values means that the filament have to be more straight, where 1 means a perfect straight line.

Window size:

The number of line points which are used to estimate the straightness. Larger values lead to more robustness against noisebut decrease to ability to detect small curves.


Allowed box overlap:

Relative amount to what extend two boxes auf different filaments are allowed to overlap. The default value 0.5 which is the maximum value that ensures that a box contains only one filament

General

Step-by-Step Tutorial