Detection masses in digital mammography images using neural networks - Thesis Example

Research Proposal Topic of Thesis: ‘Detecting masses in digital mammography images using neural networks’ Objective The basic objective of the thesisis to detect masses in digital mammography images using neural networks. I will develop a hybrid system that will also include the advantages of previous CAD systems. The new system will be able to mark breast abnormalities on digital and film mammograms more effectively than the previous CAD systems.BackgroundA group of cancer cells starts growing into the surrounding breast tissues causing a malignant tumor, which is known as breast cancer.

About one out of every eight females faces this disease. Mammography is a breast imaging technology that is used to detect breast cancer. In film screen mammography, special films and intensifying screens are used to detect breast cancer. FSM provides high quality images at low radiation doses (DeFelice 2002, p. 12). Denise and Farleigh (2005) assert, “The major limitation of traditional mammography is that the film serves simultaneously as the image receptor, display medium, and long term storage medium for the image”.

Digital mammography makes use of solid-state detectors in order to display images of breasts on the computer screen. Denise and Farleigh (2005) found that separation of image acquisition, image processing, and display to be one of the principal advantages of digital imaging system. Digital mammography also makes use of CAD (Computer-Aided Detection), which assists the physicians in image interpretation. Mass detection in mammograms refers to the detection of those groups of cells that cause breast cancer.

Bick and Diekmann (2010, p.100) found that sensitivity to be not high enough in mass detection. Computer-aided detection system, cellular neural networks, a two-stage hybrid classification network, and some other techniques can be used for mass detection. Bruynooghe (2006), in an article, found that in case of hybrid network, an unsupervised classifier is used to examine suspicious opacities, and then some supervised interpretation rules are applied to reduce false detections. Cellular neural networks play a vital role in mass detection.

Kupinski and Giger (2002) showed in a research that features extracted from potential lesion areas are sent through a neural network to decide whether the area is a true lesion or a false detection. Using CAD as a system for image interpretation is very facilitating for the physicians. However, some researchers suggest improvements in the current CAD technology. One of those suggestions includes development of a CAD system with increased ability to detect actual abnormalities instead of marking a high number of normal areas as abnormalities.

Another suggestion is to develop low cost CAD software, which should be able to study both digital images and mammogram films.MethodologyQuestionnaires and interviews with the physicians will be conducted. I will visit different breast imaging centers in order to analyze the efficiency of current CAD systems. The process of research that also includes collection and organization of data will take around one month. ConclusionSumming it up, mammography is a type of imaging technology that is used to detect various breast-related diseases including breast cancer.

Film screen mammography and digital mammography are the most widely used techniques for detecting breast cancer. Digital mammography makes use of CAD (Computer-Aided Detection), which assists the physicians in image interpretation. Along with many advantages of CAD software, current features of CAD also have some limitations that give rise to a need to develop a hybrid system that should also include the advantages of the currents Computer-Aided Detection systems.ReferencesBick, U. & Diekmann, F., 2010. Digital Mammography.

Berlin: Springer-Verlag.Bruynooghe, M., 2006. Mammographic Mass Detection using Unsupervised Clustering in Synergy with a Parcimonious Supervised Rule-Based Classifier. Digital Mammography, 40(46), pp.68-75.DeFelice, E., 2002. Breast Cancer: Web Resource Guide for Consumers, Healthcare Providers, Patients, and Physicians. Lincoln: Writers Club Press.Denise, D. & Farleigh, M., 2005. Advantages of Digital Mammography [Online] Available at: http://www.providence.org/Alaska/MedStaff/digitalmammography.

htm [Accessed 02 October 2010.]Kupinski, M. & Giger, M., 2002. Investigation of regularized neural networks for the computerized detection of mass lesions in digital mammograms. Engineering in Medicine and Biology Society, 3(30), pp.1336-39.