The US Munitions Directorate has joined in the fight against cancer by applying image-processing technology developed for target recognition in smart weapons to mammogram enhancement techniques, such as the one from the University of Florida.
Subtle details from mammograms often don’t show up because of the nature of breast tissue, particularly in younger women and women on hormone therapy.
The University of Florida’s method for enhancing mammograms is called a wavelet transform technique. In this experimental process, mammograms are scanned electronically and then converted into digitised, numerical information. The transform technique can then be executed to produce a refined image revealing subtleties that can be missed on X-ray film.
According to researchers at the University of Florida, radiologists now see only about 3 percent of the information captured on X-ray film, although these techniques may rectify the situation. With the wavelet transform technique, tumours as small as one millimetre in size can be detected.
An image processing language known as image algebra, also developed at the University of Florida, is being used to facilitate the execution of the wavelet transform technique and maximise its computational efficiency.
The image algebra is said to be capable of describing all image to image transformations, making it useful for other mammogram enhancement processing techniques. The goal of the Munitions Directorate is to provide a technology capable of evaluating all such image processing techniques in order to determine which ones perform best under various conditions.
A technological challenge with these image-processing techniques, however, is that the process is very time consuming. Using advanced workstation equipment, processing for just one parameter or specialised function of a program in digitised mammography may take from one to several hours, depending on the procedure and the diversity of features needed.
But according to theoretical studies at the Munitions Directorate, dual use technology may be able to meet that challenge.
Over the past four years the Munitions Directorate has designed a high-speed computer capable of executing all Image Algebra operations.
The design will allow computation speeds to exceed two trillion operations per second. The different parts are formulated algebraically, resulting in the capability to provide solutions to problems, such as digital mammography, enhancement in seconds.