Xray Image Editing Mini-Workshop
Images sent from one orthopaedic surgeon to another or
to the email list need to have a small file size, but above all, they need to convey the
information about the case. This conflict between images which are inadequate and files
which are too large can only be resolved if the senders pay some attention to editing the
original image. This workshop lays out the stages between the original scan or digital
photograph of the Xray and the image finally edited for sending. It takes very little time
and may be the difference between success and failure if you are asking for help. This
page uses a popular Image Editing program Lview which
costs about $40 USD and has some more sophisticated features. The companion
page uses IrfanView a freeware image editor (no
cost). There are literally dozens of freeware image editing programs, many of which are
specifically designed for use with Xray images. A page called Introduction to the
DICOM standard and free software explains the Digital Imaging and Communications in Medicine (DICOM) standard created
by the National Electrical Manufacturers Association (NEMA) to aid the distribution and
viewing of medical images. This page also has links to a wide variety of DICOM image
editors. OSIRIS
DICOM Viewer was recommended to the Orthopod Mailing list by Alex Chelnokov. It has the
advantage that one can use measuring tools etc.
In many situations the original Xray is not an adequate
study with non-standard views or poorly exposed films. This document cannot address that
issue beyond saying that sending the patient back until views of the required quality are
obtained, may eventually convince Xray departments that it is cheaper and easier to do it
right first time! We will assume we have to make do with the Xray films we have. Much of
this workshop also can apply to digital radiographs which are computer files from the
beginning.
Here is the initial image sent in by Hani al-Elesh.
It's size is 19.5 kb.
Particularly as far as image size is concerned this is much better
than many images sent to the list which tend to be much too small in area. The image has
also been converted to grayscale instead of colour which eliminates the greenish cast from
the fluorescent tubes in the Xray viewing box. That said the image does not show much
detail in the critical areas of the both bones fracture and the elbow region. Most image
editing programs, including some free ones, have the tools that are required to improve
this image. I have used the Lview Pro image editor available in free evaluation form from http://www.lview.com ($40 USD for purchase)
- Step 1. Making the original scan/digital image
You want as much detail as possible in both the over and the underpenetrated
areas of the Xray. Until you are very happy with your technique it would be a good idea to
take one picture 2 stops overexposed (detail in the dark areas) and one picture 2 stops
underexposed (detail in the light areas). Chose the one which is best for the part of the
image which is important to you.
- Step 2. Convert to Grayscale
This has already been done for al-Elesh's image so I am illustrating it with one
of mine. In LView it is done in the Color/Adjustment Menu. 
Select Grayscale. Note that the original seen to the left has a
blue/green tint. Once the Grayscale option has been selected the thumbnail sized image is
shown monochrome. Once the "Apply" button is clicked the main image changes to
monochrome. (Note this may be different in other image editors)
- Step 3 Crop and Tilt You only need to send the parts
of the image that convey information. The less empty space in the picture the larger it
can be for the same overall file size. In this instance we are only interested in the
elbow and forearm; the other injuries are not illustrated in this film. By angling the
cropping rectangle we can make it smaller.
Once cropped the image "straightens out"
- Step 4 Resize
Use the Image/Resize menu if the image is too large for the screen. In general
600X400 pixels for landscape and 400X600 for portrait orientation is about right. Any
bigger and the image may be too big for the screen; any smaller and the orthopaedic
content may be lost. There is not much point in increasing the size of an image as you
don't generate any more information. In this case the image as sent was a good size but
many others sent recently have been too small.
- Step 5 Use Histogram Function to improve contrast
and brightness
This function analyses the amount of dark and light pixels in the image and produces a
frquency ditribution histogram. In this instance it shows the mode (arrow) to be displaced
towards the left hand end of the distribution (the dark side!)
The histogram also shows that although there are some very dark pixels
there are no very light pixels.Therefore we will not lose any information by moving the
upper limit (white triangle) leftward to the right hand end of the distribution. This has
the effect of making the lighter pixels even lighter so that some of them are nearly
white.
The elbow fracture which was already quite light is shown better but it
still doesn't help to display the shaft fractures. To do this we need to lighten up the
middle tones. That is done by dragging the mid tone arrow (gray) leftward until it is
nearly at the mode. What this does is "spread out" the lower end of the
distribution and compress the upper end resulting in lightening up of the mid tones.
The end result is a smaller image (14.1 kb) but one with nearly all the
information the original Xray contained.
You can do much of this manipulation with brightness and contrast controls, or with the
Curves tool but in my opinion it is easier with histograms for the specific case of
orthopaedic Xray images. Once you are familiar with the way your editing program does
this, all of it can be done in less than a minute.
Compare the original
Comments and questions to the Orthopod Mailing List please
Myles Clough mylesclough@shaw.ca
July 11th 2005
Image Editing using IrfanView