Automated Analysis of Mammograms using Evolutionary Algorithms

Breast cancer is the leading cause of death in women in the western countries. The diagnosis of breast cancer at the earlier stage may be particularly important since it provides early treatment, this will decreases the chance of cancer spreading and increase the survival rates. The hard work is the early detection of any tissues abnormal and confirmation of their cancerous natures. In additionally, finding abnormal on very early stage can also affected by poor quality of image and other problems that might show on a mammogram.
Mammograms are high resolution x-rays of the breast that are widely used to screen for cancer in women. This report describes the stages of development of a novel representation of Cartesian Genetic programming as part of a computer aided diagnosis system. Specifically, this work is concerned with automated recognition of microcalcifications, one of the key structures used to identify cancer. Results are presented for the application of the proposed algorithm to a number of mammogram sections taken from the Lawrence Livermore National Laboratory Database.
The performance of any algorithm such as evolutionary algorithm is only good as the data it is trained on. More specifically, the class represented in the training data must consist of the true examples or else reliable classifications. Considering the difficulties in obtaining a previously constructed database, there is a new database has been construct to avoiding pitfalls and lead on the novel evolutional algorithm Multi-chromosome Cartesian genetic programming the success on classification of microcalcifications in mammograms.