Finding suspicious masses of breast cancer in mammography images using particle swarm algorithm and its classification using fuzzy methods

Mammography images which are one of the latest methods of breast imaging can largely help in detecting tumors. But, because of error possibility in determining benign and malignant tumors by physicians, an intelligent system for interpreting these images and diagnosing calcium tissues can always prevent from unnecessary biopsy and human visual errors. The objective of this article is this and there is an attempt to use wavelet transform to extract image features; then, using the particle swarm algorithm, the features which were more important and effective were selected. Finally, the obtained results were converted to fuzzy rules, an inference was made of these rules and the images were diagnosed and classified. The results were tested on MIAS database and the accuracy of 93.41 percent was obtained. These results were compared with those of three others. Also, the criteria of sensitivity and specificity were improved.