Proceedings of The 2nd International Conference On Advance And Scientific Innovation, ICASI 2019, 18 July, Banda Aceh, Indonesia

Research Article

Feature Extraction Performance Verification based on Ultra-wideband Imaging and Artificial Neural Network

Download284 downloads
  • @INPROCEEDINGS{10.4108/eai.18-7-2019.2288670,
        author={V  Vijayasarveswari and M  Jusoh and T  Sabapathy and M N Osman and R.A.A.  Raof and S  Khatun and M N Yaasin},
        title={Feature Extraction Performance Verification based on Ultra-wideband Imaging and Artificial Neural Network},
        proceedings={Proceedings of The 2nd International Conference On Advance And Scientific Innovation, ICASI 2019, 18 July, Banda Aceh,  Indonesia},
        publisher={EAI},
        proceedings_a={ICASI},
        year={2019},
        month={11},
        keywords={feature selection feed forward backpropagation neural network signal processing},
        doi={10.4108/eai.18-7-2019.2288670}
    }
    
  • V Vijayasarveswari
    M Jusoh
    T Sabapathy
    M N Osman
    R.A.A. Raof
    S Khatun
    M N Yaasin
    Year: 2019
    Feature Extraction Performance Verification based on Ultra-wideband Imaging and Artificial Neural Network
    ICASI
    EAI
    DOI: 10.4108/eai.18-7-2019.2288670
V Vijayasarveswari1, M Jusoh1,*, T Sabapathy1, M N Osman1, R.A.A. Raof1, S Khatun2, M N Yaasin1
  • 1: Bioelectromagnetics Research Grant (BioEM), School of Computer and Communication Engineering, University Malaysia Perlis, Perlis, Malaysia
  • 2: Faculty of Electrical & Electronic Engineering, Universiti Malaysia Pahang, Pekan, Pahang, Malaysia
*Contact email: muzammil@unimap.edu.my

Abstract

Breast cancer is a second leading case among the women. Therefore, an efficiency breast cancer system is very important for accurate early detection. The selection of number of features is very important to avoid the system complexity and large processing time. This paper proposed a modified feature selection method for early breast cancer detection. Ultra-wideband (UWB) signals are transmitted and received using a pair of antenna. 1632 features are extracted from the received UWB signal and four statistical features (mean, median, maximum and minimum numbers) are selected from the extracted 1632 features. These features are fed into feed forward backpropagation neural network for breast cancer detection. The proposed features selection method is able to use to detect the breast cancer in terms of existence, location and size with average accuracy of 86.28%. The proposed feature selection method is able to increase the system performance