Digital Mammography X-ray Image Enhancement And Detection Of The Tumor Region

Breast cancer remains a leading cause of cancer deaths among women in many parts of the world. There are many tools to examine the breast cancer in clinical diagnosis. Mammography is an effective technology to checking out the early breast cancer. Masses in a mammogram have much meaning to find an early indication of breast cancer. Mass is an important symptom of breast cancer that usually likes a rounded convex region. In the mammograms the contrasts between the masses and their backgrounds are commonly very weak, and the boundaries of mammographic masses are often very vague. However, the boundaries characteristics of the masses are the important clues to discriminating malignant and benign tumors. So it is difficult for doctors to detect the masses correctly and quickly. At present, there are so little reports about the research on the enhancement and detection of mammographic masses. This paper proposes some new methods that are verified effectively in our experiments include of mass enhancement, mass segmentation, mass recognition, display consistently, and so on.Mass enhancement: Because there are small differences in X-ray attenuation between normal glandular and malignant tissue, mass is hard to be detected for presenting low local contrast, and the boundaries of the masses are usually vague. Based on the features of mass, a better method that can enhance and detect the mass areas effectively was proposed in this paper. By using the gradient information of the image, this method can enhance the mass even if the contrasts to their background are very low. Meanwhile the boundaries features of the enhanced object regions were held.Mass segmentation: The shape characteristics of mass are important clues in discriminating malignant and benign tumors. A new region growing with four parameters including of gray, area, contrast, and compactness, those are extracted from the enhanced images,can be used to segment the mass region effectively and the edges features of mass are keep well by processing the whole mammogram.Mass recognition: A rough sets theory is a new mathematical tool to deal with uncertainty. It has been widely used in the area of machine learning, data mining, pattern recognition, etc. This paper proposes a mass recognition system that based on rough set. This system contains two parts rule acquisition and mass recognition, both are related to rough sets. Rule acquisition contains several steps: image pre-processing, mass segmentation, feature acquisivertion, decision table establishment, decision table discretization, decision table reduction, rule acquisition. Masses are recognised by matching rule. The result shows that this method can discriminate malignant and benign tumors effectively.Display consistently: Because different display systems have different characteristics, the same image can't display the same effect in different display systems. However, the display effect of the medical image is very important to diagnose the illness. This paper provides a method to modify the display effect of medical image,based on human visual property. Not only the display effect of the image can be improved, but the image can be also displayed consistently on different systems by using this method.Using VC software, we combine tumor detection with a mammogram database management system to implement a computer-aided mammogram management and diagnosis system. Such system will have broad prospects.