More Deconvolution Options

Field

Description...

Prefilter Resolution Limits

Resolution limits (two numbers) for the Gaussian pre-filter, measured in inverse microns. You should not change this value. The Gaussian smoothing curve starts at the first number, drops by 3 standard deviations at the second number, and is 0 thereafter. Using 0.0 and 0.0 turns this filtering step off. The default values are 0.30 and 0.22.

Smoothing

The amount of smoothing to apply to the image between iterations. The default value of 0.2 is usually adequate.

Border Rolloff (voxels)

The width of the border rolloff (in voxels). For images with bright spots around the border you may want to increase the border rolloff. To minimize edge effects, the border rolloff is automatically set to about 1.5% of the image dimensions.

Crop Border Rolloff
after processing

Remove specified border rolloff after the image has been processed.

Size for Z Transforms

The size of the Fourier Transform along the Z-axis. Increasing the Z Transform size can occasionally improve deconvolution results. The default is the power of two that is at least 1.3 times larger than the size of the image along the Z-axis.

Wiener Filter Enhancement (0-1)

The amount of Wiener filter enhancement. This applies only to enhanced ratio deconvolution. The default value of 0.9 is usually adequate.

Wiener Filter Smoothing (0-1)

The amount of Wiener filter smoothing. This applies only to enhanced ratio deconvolution. The default value of 0.8 is usually adequate.

Intensity Scale Factor

An intensity scaling factor that is applied to the images during deconvolution. Scaling is usually unnecessary because the deconvolution program automatically scales images to fit the dynamic range of the resulting image.

Field

Description...

Normalize Intensity

Control whether to use intensity normalization for the image correction. The default is on.

Use Photosensor

Control whether to use photosensor readings for intensity normalization. The default is on.

If this button is off and the normalize correction button is on, the image correction program normalizes intensities based on a best-fit polynomial, rather than photosensor readings. The best-fit polynomial method is a reliable alternative to the photosensor method.

Note:  If an image contains invalid photosensor readings, the image correction program automatically switches to the polynomial method. In this case, the switch is recorded in the corresponding "log" file.

Correct Bleaching

Correct photobleaching. For situations where image intensity changes are due to the specimen and not photobleaching, turn this option off. The default is on.

Replace Z-Lines

Use the Z-Line correction. This is designed to detect CCD defects by searching for unusual intensity profiles along the Z-axis. For example, a "dead" CCD element appears as a straight, dark line along the Z-axis, and a "hot" CCD element appears as a straight, bright line along the Z-axis. Voxels identified as Z-line errors are replaced with the average of their neighbors. The default is on.

Smooth Z-Lines

Smooth the stack along the Z axis. The default is on.

Camera Intensity Offset

Specify a Camera Intensity Offset parameter. Conventional CCD cameras will exhibit an offset in the approximate range of 50-100 counts, whereas EMCCD cameras may be in the many hundreds. This value can be measured by acquiring a bias image at the target camera gain and noting the image mean value.

Note:  The bias image will change with the particular experiment variables, such as

binning, exposure, and camera gain. Therefore, it is essential that you measure the

Camera Intensity Offset under the same conditions in which you are imaging.

Pass Waves Unprocessed

Passes data for selected wavelengths through without applying any correction or deconvolution processing. This is useful if a transmitted-light channel is collected along with a fluorescent channel.

Note:  You can set a system configuration variable (DV_PASS_WAVES) to specify a list of wavelengths to be ignored by default. You can set the variable locally in the .dvrc file in your home directory, or globally (for the entire system) in the system.dvrc file in the /usr/local/dv2.00/config/ directory. For example, you may arbitrarily assign the wavelengths 100 and 101 to be two types of transmitted-light imaging as follows:

DV_PASS_WAVES 100 101

Subtract

A simple background subtraction. The default is 0. Note that excessive subtraction can cause deconvolution holes.

Save Intermediate Results as 2-byte Integer

Always saves intermediate images as 2-byte integers. In some cases, the deconvolution may save intermediate images as 4-byte floating-point numbers.  

Save Final Results as Floating-point

Saves results as floating-point. (Deconvolution images are calculated with floating-point numbers.) In most cases, 2-byte integers have sufficient precision and magnitude to represent fluorescent images acquired with 12-bit CCD cameras.

Maximize Use Of Memory For Speed

Increases the deconvolution speed. This is useful when you are deconvolving relatively small images and have a large amount of physical memory in your workstation. When this option is used, interrupting the deconvolution process may result in an unusable output image.

Automatically Terminate

Automatically terminates deconvolution when the deconvolution residuals grow larger for two consecutive iterations. This option is selected by default. If the deconvolution process returns to convergence after the residuals grow, try deselecting this option.