Indian Journal of Science and Technology
DOI: 10.17485/ijst/2017/v10i3/110640
Year: 2017, Volume: 10, Issue: 3, Pages: 1-6
Review Article
Susama Bagchi and Audrey Huong*
Universiti Tun Hussein Onn Malaysia, Malaysia; [email protected], [email protected]
*Author for correspondence
Audrey Huong
Universiti Tun Hussein Onn Malaysia, Malaysia;
[email protected]
This work aimed to review different Computer-Aided Detection (CAD) systems that were proposed as alternative means to replace the tedious and erroneous double reading procedure via radiologists. These CAD systems include the use of different signal processing techniques such as Wavelet Transform and Curvelet Transform, image processing technique, pattern recognition, artificial intelligence technologies namely the Artificial Neural Network and Fuzzy Logic system, and different algorithms of computer sciences. Among the developed algorithms proposed for this purpose include k-nn algorithm, fuzzy C-means (FCM), swarm algorithm, genetic algorithm and multi-resolution techniques. It was found from this study that multiscale curvelet transform has the highest classification accuracy with the reported value up to 98.59 %, followed by the Swarm Optimization which produced a percentage error of 1.7 %. Meanwhile it was observed that multi-resolution technique along with genetic algorithm produced the highest error of 20.8 ± 8 % in its diagnosis. This work concluded that curvelet transform and swarm algorithm is, thus far, the most suitable CAD techniques to be used before clinical investigation of malignant breast tissues. In the future, these techniques may be improved further to detect different stages of breast cancer.
Keywords: Breast Cancer Detection, Computer-Aided Detection System
Subscribe now for latest articles and news.