원문정보
초록
영어
This paper, presents a competent feature extraction technique, i.e., Bidimensional Empirical Mode Decomposition (BEMD) for mammogram images. The EMD is fully adaptive and data driven technique. BEMD is used to extract features at numerous scales or spatial frequencies in form of Intrinsic Mode Functions (IMFs). By using these IMFs five statistical textural features i.e., mean, standard deviation, kurtosis, skewness and entropy have been extracted from Region of Interests (ROIs) of preprocessed digital mammograms. Artificial Neural Networks (ANN) which is inspired by biological neurons has been explored to classify mammograms into different classes. Three experiments, i) classification of normal mammogram and calcification, ii) classification of mass tissue and calcification and iii) classification of normal, mass and calcification have been performed. Accuracies of 95.5%, 93.2% and 82.4% have been obtained by proposed method from respective experiments.
목차
1. Introduction
2. Literature Review
3. Overview of Bidimensional Empirical Mode Decomposition Method
4. Methodology
4.1. Digital Mammogram Database
4.2. Pre-processing
4.3. Region of Interest Selection
4.4. Feature Extraction using BEMD
4.5. ANN Classifier
5. Experimental Results
6. Conclusion
References