Research Of Mammogram Enhancement Algorithms

Breast cancer is one of the most frequent malignant women diseases.Mammography is one of the effective ways for early detecting the disease.Because of the limitations of the X-ray hardware systems and uneven tissuethickness, conventional mammography suffers several drawbacks such aslow resolution and low contrastThus, in order to improve the correct diagnosis rate of cancer, imageenhancement techniques are often used to enhance the mammogram andassist radiologists in detecting it. Histogram equalization is a widely usedtechnique for image enhancement. After a conversion to a uniform gray leveldistribution, several minor grey levels would merge into only one grey level,resulting in a lower portion of the human eye gray distribution. Moreimportantly, the details included in the minor grey levels will be lost afterimage equalization although equalization itself is supposed to make details.In this paper, we will investigate the image enhancement technologybased on background homogenization. The steps of the method can bedivided into luminance compression and histogram equalization. Throughthe luminance compression, bright areas become dark and dark areasbecome light, i.e. luminance range gets smaller. After applied by histogramequalization, the luminance difference increases, and the dark detailsbecome very apparent. Local contrast enhancement are used for furtherenhancement. Our methods can effectively improve the visual quality ofmammograms by enhancing the contrast, especially highlight the fine detailsof dark regions in mammograms. Doctors diagnose the illness by viewing medical images presentedinformation. Therefore, the results of breast X-ray image enhancementshould meet the human visual system. According to the visual characteristics,the image is first broken up into the different regions of human visualresponse. These different regions are characterized by the minimumdifference between two pixel intensities for the human visual system toregister a difference. Next, these three regions are thresholded, removing thepixels which do not constitute a noticeable change for a human observer andplacing these in a fourth image. Experimental results demonstrate thatproposed algorithm is effective for the enhancement of breast X-ray images.