XTAtrous

XTAtrous for filtering (2-D / 3-D) image stacks

XTAtrous is a convenient wrapper around the libatrous library, which interfaces with Imaris for filtering 2-D / 3-D image stacks (and time series).

Its main use is for pre-filtering datasets prior to spot / volume tracking.

The filter parameters for each channels are automatically embedded in the channel description (JSON object) but can also be manually exported to a file for easy reuse.

I have documented each of the parameters and option via tooltips. Just hover on top of the labels to display them.

Parameters

• Channel: Which channel to analyse. A new channel will be created for each filtered channel.
• Filter type: Any of 'Linear 3x3', 'B3 Spline 5x5', 'Cubic alpha=-1', 'Cubic alpha=-0.5', 'CDF 9/7 (JPEG 2000)', 'Gaussian 5x5', 'Linear 5x5'
  • Choose the 3x3 to preserve small features.
  • The 'B3 Spline 5x5' and 'CDF 9/7' are interesting for tracking cells and larger objects.
• Lower limit: The first low number scale to allow through the filter. If you have a noisy image, you may want to disregard scale 1 and possibly scale 2.
• Upper limit: The largest number scale to allow through the filter. To create a band-pass filter, set this to be the same scale as the lower limit scale. For example, to detect spots with diameters around the 16 to 20 pixel mark, try using the 3x3 kernel and set both lower and upper limits to 4.
• Options:
  • Check the threshold box to use the sliders below to clip the low / high pixel intensity values. In Threshold only (no normalisation) mode, pixels whose intensity is below the minimum threshold are set to 0.
  • Check the normalise box to scale the output intensity range to that of the input image. For example, if the input image is 16 bits, the normalised output image range will span 0 to 65535. For quantitative comparisons where pixel intensity is important, you will want to work with non-normalised images.
  • Check the low pass residual box to add the residual low-pass filter to the output (generally not needed as you will most certainly try to remove the background).
• Threshold min: The low threshold limit clips low intensity / negative intensity pixels.
• Threshold max: This threshold clips high intensity pixels.
• Preview: Click preview to filter a data set for the current timepoint and display the filtered data in Imaris. Filtered data is cached so that no calculations are performed unnecessarily. Also, click preview to update the filtered image's pixel intensity range.
• Apply: Click this filter every timepoint in your image/volume timelapse (this may take sone time).

Example (Algae dataset)