Help for the IDP3 Pixel Repair Function

   Two methods of bad pixel repair are available: gaussian interpolation
and box median.  Bad pixels may be identified in any of six ways:
    
    1. from an ascii list of x,y pixel positions specified in the Bad Pixel
       List field.
    2. from a mask fits file specified in the Bad Pixel Mask field where 
       the bad pixel value in the mask is identified in the Bad Pixel Value
       in Mask field.  If an internal mask has been loaded for the current
       image, it may be selected with the Use Internal Mask switch.
    3. from the identification of all pixels with a given value specified
       in the Value field.
    4. from the identification of all pixels greater than a given value
       specified in the Greater Than field.
    5. from the identification of all pixels less than a given value
       specified in the Less Than field.
    6. from the list of NaN values that were identified when the image
       was loaded.

Multiple methods may be specified simultaneously and the results of each
method are combined in a final mask for repairing data.  The Gaussian
Interpolation repair method requires an integer Kernel Diameter specied
in pixels (should be set to fix(0.5 + 2.2*lambda/D) and the Full Width 
Half Maximum.  Nominal values for Camera 2 F160W  radius and fwhm are 
5 and 2.2.  The repair algorithm was written by Inseok Song and Glenn
Schneider.  The Neighbor (Box) Median requires a Box Size specified in
pixels.  The box must be square.  If more than 2 pixels in the box
centered on the current bad pixel are bad no repair is made and an
error is reported.  If two pixels within the box are bad a median is
computed excluding the second bad pixel.  When the second bad pixel
is selected to be repaired the first bad pixel is excluded from the
median computation.

The final repaired image may be saved to memory and/or disk.  When a
single image is repaired (only one image is ON) the user may specify
a new filename or an append string to be applied to the original file-
name in the Append field for the output file.  When multiple images
are ON, the user must supply an Append field which is applied to the
original input filenames.

The repair is always applied to the original image as loaded into idp3.
If the repaired image/s are saved to disk any shifts, rotations, etc.
that have been applied to the image/s during the current idp3 session
are lost.  If the image/s are saved to memory (the idp3 images structure)
all shifts, rotations, etc are retained in the new image entry in the 
images structure.  All of the specified operations will be applied to 
the data before it is displayed.