Early Breast Tumor Detection Using NN-Based UWB Imaging

摘要：

Achieving successful breast cancer treatment is highly dependent on early detection. The use of feed-forward back-propagation Neural Network (NN) model to detect as small as 100 μ-meters in diameter is presented in this paper. The NN is trained, validated and tested on feature vector data constructed from Ultra-Wide Band (UWB) signals. The UWB signals at 4 GHz (center frequency) are transmitted from one side of a 3-D breast model and received on the other side. The 3-D breast and tumor models were constructed using Electromagnetic (EM) simulator. Very promising results (about 100% and 93% single tumor presence and size detection rate respectively) have been achieved.

作者: Saleh A.Alshehri Sabira Khatun Adznan B.Jantan Raja Syamsu Azmir Raja Abdullah Rozi Mahmood Zaiki Awang

作者单位: Department of Computer and Communication Systems Engineering, Faculty of Engineering Department of Computer Systems and Networks, Faculty of Computer Systems and Software Engineering Un Department of Imaging, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, 43400, Se Department of Electrical Engineering, Faculty of Engineering Universiti Teknologi Mara

会议类型: 国际会议

会议名称: 2010国际超宽带会议(ICUWB 2010)

会议地点: 南京

会议语种:英文

页码: 1-4

在线出版日期: 2010-09-20（万方平台首次上网日期，不代表论文的发表时间）