Sensor Shootout Test Explanation
This comparison test was undertaken to provide an unbiased way for practitioners to compare the performance of currently available direct digital (DR) dental X-ray systems commonly used in the veterinary marketplace. Any system is capable of producing fantastic images at a tradeshow booth. This may or may not be representative of the quality of images you will view with your own clinical patients. The only way to truly compare systems is to have a side-by-side comparison available for viewing. A short explanation of the test follows. VetDentalRad.com is just providing images for you to use for comparative purposes. We strongly encourage you to investigate software issues and company support prior to purchasing a digital dental radiography system. Several references from long-term users should be contacted as well. We hope that you enjoy the images.
The four views in this comparison were selected for specific reasons:
1. Canine lower first molar using a standardized parallel technique: This view was selected to show the image quality obtained with the beam directly aimed at (perpendicular to) the sensor.
2. A canine upper fourth premolar, utilizing a standardized 45 degree bisecting angle technique: This view was selected for several reasons. Since upper fourth premolars are commonly radiographed, this view is often used in veterinary dentistry. This view also tests the ability of the sensor to function with a non-parallel incident angle, also known as a “bisecting angle” technique. Some sensors are more sensitive to the incident angle between the beam and sensor than others, with resultant degradation of the images as the beam becomes less perpendicular to the sensor. Since most dental imaging in veterinary patients require a non-parallel (bisecting angle) technique, this is an important factor to consider.
3. Feline maxillary premolars using a standardized “near-parallel technique”: This view, perhaps the most challenging view for any digital system, was chosen for two reasons. It provides a test for a sensor using a non-parallel technique as described in #1 above. Additionally, this view contains a wide range of densities, ranging from air to the thicker portion of the Zygomatic arch seen in feline maxillary radiographs. Different digital radiography systems vary in their ability to handle a wide range of densities when present on the same image. The software algorithms present in the software have a profound influence on how well wide density ranges are handled and presented on the resultant image.
4. Line-pair grid: The line-pair grid gives a rough measure of the sensor’s resolution. Many manufacturers publish figures for “theoretical line pairs per mm”. This is meaningless. Theoretical LP per mm is an esoteric figure arrived at through mathematical calculation. What difference does a theoretical LP measurement make if you will never see that degree of resolution when you use the product? These images were taken with a line-pair grid, which has multiple fine wires that slowly converge from one end of the grid to the other. The numbered scale at the bottom of the images refers to the actual number of line-pairs per mm present at that point on the image. As the lines converge, they become less distinct, finally blending into a blur. When viewing the line-pair images, try to observe the point at which you lose clear resolution between the lines. That is the “actual or observed” line pairs per mm. You may want to magnify these images for a better idea about resolution. All of these systems ranged in resolution between around 6 and 15 line pairs per mm. Line pair ratings are only one part of the total picture. Software has a major impact on the quality of image. All other things being equal, it is more desirable to have better resolution, but resolution is not the only measure of a system’s performance.
The selected images were not manipulated after they were taken. All images were exported as JPEG images. When using JPEG compression, there is some loss of detail in all of the images. Please note that the Schick and vetX systems export JPEG files of a smaller size than the other systems. This means that image quality of the Schick and vetX systems were probably more adversely affected by choosing this format for the files. This effect will be most apparent if the images are magnified. This disparity in file size will be eliminated next year when the DICOM 3.0 format will be utilized for this test. All of the digital dental systems have a large variety of enhancements that can be used to manipulate the images after they are viewed on the screen. None of the images presented here were manipulated after they were on the screen.
Although DICOM 3.0 is generally considered the future standard for all medical imaging, including dental radiography, it is recognized that most digital dental systems are currently not capable of exporting images in this format. Next year, we will be evaluating all the systems for DICOM 3.0 compatibility. Such compatibility radically simplifies sharing digital images in a secure, standardized, tagged format.
If anyone has any suggestions for future tests, please
feel free to contact us at admin@vetdentalrad.com.
2/15/07
Brook Niemiec DVM, AVDC, FAVD