Mammography Detection System of Malignant breast mass Cancer Using Hybrid Expert System and Case Based Reasoning

soran Saeed, Bestan Maaroof, Alla Shally
2018 Evaluation Study of Three Diagnostic Methods for Helicobacter pylori Infection  
This research work, presents a computer-aided mammography detection of mass image for Malignant breast cancer a system has been developed to help radiologists in order to increase diagnostic accuracy and called (ImageCBR). The aim of this work to find or detect similar Malignant image mass of breast cancer from base knowledge by given a target one. similarity Generally, a ImageCBR system consists of four stages: (a) preprocessing of the image (b) segmentation of regions of interest, such as a
more » ... ll-known mass breast features extraction and selection (shape, size, density, margin), and finally (c) image similarity (target and source). The performance evaluation metrics of ImageCBR systems are also reviewed. Figure 2.0 The structure of the breast and breast cancer. Architecture design Figure2.0 shows the architecture design of the proposed system. The integrated expert system and image processing attempt to increase the accuracy of the casebased reasoning detection of mammography breast cancer obtained from classical Case-based reasoning. When a new case is arriving the image is pre-processed and analysed by the image processing and the image is re-analysed by our case doctor features are extracted and stored in the base knowledge. The CBR cycle [7] starts when the new case (target case) is arrived; in the RETRIEVE step all cases in the system are retrieved that are similar to the new case. Then REUSE the solution of the most similar case is Reused for the new case, in the REVISED step the suggested solution is revised or tested whether it is suitable for the new case or not. Finally the test case is added to the database in the RETAIN step as a new learned case in this way the CBR will get more and more clever as a human expert as he sees more cases he gets more clever.
doi:10.24271/garmian.302 fatcat:ugmz6l63gzenjk4c7avik5wy64