PCA Based Dimensional Data Reduction and Segmentation for DICOM Images

Digital Imaging and Communications in Medicine (DICOM) is a trendy for a clinical picture area. The modem tools for the acquisition of pixel has a DICOM interface, which lets in interoperability between gear and the storage in documents of the equal format. However, it is applicable that the pictures and its associated statistics be built-in to the Hospital Information System. The image captured through the capturing devices contains features with higher dimension which increases the time complexity in classification. As the classification performance greatly impacts on the diagnosis of different diseases and supporting the medical practitioner, it is necessary to reduce the features size or dimensionality. Also, by reducing the dimensionality of features, the decisive support system would provide more accurate result to the medical practitioner in making different decisions. The existing techniques suffer to identify the exact features by segmenting the features according to different features. The computerization of a hospital, in a built-in manner, is now not a convenient task. Developing statistics structures that can combine scientific photo records in to the different hospital’s information, except growing interfaces, which can compromise the system’s performance, is critical to add such traits in the facts device itself. Rand index of the image, Global Consistency Error, Variation of Information and PSNR are major scalable parameters in the various segmentation methods. The mentioned parameters are needed to enhance significantly for the better detection of an image through segmentation process. Hence, principle component analysis method and expectancy maximisation based segmentation methods are proposed in this effort.