Research On Digital X-Ray Medical Image Enhancement Methods Based On Human Visual Properties

Degradation of digital X-ray medical image such as low contrast and blurring during radiographic imaging, caused by complexity of body tissue and effects of X-ray scattering and electrical noise etc., can worsen the results of analysis and diagnosis. So it is necessary to enhance X-ray medical image to improve its visual quality, including contrast enhancement and detail enhancement.Review of some existing enhancement methods for medical images points out that limitation of these methods lies in their taking no account of human visual properties and diverse fabrics propeties and density properties of different body part.In this paper, existing enhancement methods of digital X-ray medical image are improved for different body part, and image processing methods based on human visual properties are presented. The content of this paper are summarized up as follows:A modified sine gray-scale transformation method is presented, with regard to the complexity of selecting parameters of linear transformation and the limitation of nonlinear transformation such as logarithm and exponent transformation. Spatial correlation between neighborhood pixels in X-ray medical images is utilized to reduce loss of visual details.Fast algorithm for adaptive neighborhood gray-scale stretch histogram equalization method is proposed which can reduce operand greatly.An adaptive fuzzy contrast enhancement approach based on homogeneity detection is presented which uses image fuzziness and the visual property that sensitiveness to noise of human eyes is higher in area of homogeneity than in area of detail to enhance local contrast of image and suppress amplification of noise effectively.An adaptive unsharp masking approach based on region segmentation is proposed which mainly enhances areas of medium contrast and maintains areas of smoothness to avoid noise amplification in areas of smoothness and enhances areas of high contrast properly to avoid over-sharpen for edge.An adaptive contrast enhancement method based on contrast sensitivity function is presented. This method utilizes the visual property that object can be discriminated if its contrast is higher that human eye' s threshold contrast. It can ensure that areas of image detail obtain enough contrast and that the image is not over-enhanced at the same time.A multi-channel adaptive homomorphic enhancement approach is presented which uses the following visual properties: nonlinearity, multi-channel filtering and sensitivity to high frequency information. This method can enhance image details of various sizes effectively. X-ray mammagraphs processed with this method give favorable results.