Improved SMQT Algorithm and PCNN Model for Micro-calcification Clusters Detection in Mammograms

This paper proposes a novel combined method to improve micro-calcification clusters (MCs) detection accuracy in mammograms. The presented method includes three main steps: firstly, exponentiation operation and a new improved successive mean quantization transform (SMQT) algorithm are employed to enhance MCs; secondly, wavelet transform is introduced to obtain the significant MCs information; thirdly, pulse-coupled neural network (PCNN) model is used to detect MCs. In the experiment, totally 73 mammograms from MIAS database and 41 mammograms from JSMIT database are chosen to test the algorithm, and experimental results demonstrate that the algorithm presented in this paper is better than the other algorithms by yielding higher specificity of 98.0%, accuracy of 97.26%, and sensitivity of 95.65%. Besides, the method is verified on 20 mammograms from the People's Hospital of Gansu Province, and the detection results indicate that our algorithm can detect MCs correctly. Above all, the proposed method is simple and effective, and it can be considered to assist the radiologist for breast cancer diagnosis.