Deep sparse auto-encoder for computer aided pulmonary nodules CT diagnosis

In this paper, we develop a computer aided diagnosis algorithm to detect and classify the abnormalities in lung CT examination. We focus on analyzing the pulmonary nodules CT data and help the medical experts improve the accuracy of medical diagnosis with our analysis tool. To achieve this, we take into account the advantage of sparse representation in medical signal and design a novel deep sparse auto-encoder (DSAE) model based on deep learning. Unlike the other traditional feature selection models, this DSAE model can extract the underlying features of pulmonary nodules automatically and these features may a better presentation of original medical CT data. For the experiments, we use a pulmonary CT dataset and conducted experiments and comparisons with other algorithms to justify the effectiveness and efficiency of our algorithm.