Multi feature malignant weight-based mammogram classification with ANN using fuzzy rules
B. Sri Ramya, Dr. K. Arthi · International Journal of Engineering & Technology · 2019
Towards the development of disease prediction, detection and classification accuracy, the features of mammogram can be used to estimate the similarity. In this paper, we motivate to adapt the fuzzy logic with ANN to perform classification. With the motivation, a multi feature fuzzy logic analysis with ANN Based Mammogram Classification algorithm is presented. First, it is necessary to obtain the gray features of mammogram. The method considers the brightness of nodule, lobulation, granularity and nodule size. Using the features considered, the method generates the fuzzy rules. The artificial neural network has been generated with number of neurons where each layer of the neuron has been initialized with specific feature. At the testing phase, the input features have been used to estimate different weight measures using the fuzzy rule generated. The method estimates multi feature malignant weight (MFMW) using the fuzzy rules at each layer of the neurons. Finally, the classification is performed based on the value of MFMW.