Application Research Of Medical Digital Grayscale Image Processing Technique

The thesis introduces the application of medical image processing technique on three aspects from the angle of medical application, including gray scale transformation, image noise removal and image sharpening. The thesis mainly encloses the application on processing of medical CT image.Medical images include ultrasonic image, X-ray image (including CT image), nuclear magnetic resonance image, nuclear medical image, microscopic image, DSA image and various kinds of endoscopic image, etc.In recent decades, digital medical image and digital medical image processing technology developed greatly, and had formed a complete branch in medical domain, named medical image. Medical image contain abundant information of biomedical anatomy structure and function. Reading out correctly and understanding this information deeply can provide emphatic support for medical diagnosis, treatment and medical investigation.The thesis introduces the concept of image, clearly indicates that image is the distribution function f(x,y), which information is recorded on two-dimension plane. An analyzable binary function f(x,y) is the basic mathematical model of image. If the value of function f(x,y) is continuous which constitutes the characteristic of pictorial picture cell, the image is analog image. If the value of function f(x,y) is straggling, the image is digital image. This essay also introduces the basic concepts of image, such as pictorial grayscale, grayscale value, maximum grayscale value, minimum grayscale value, grayscale average value, grayscale variance and image entropy, and so on. This essay even more elucidate the concept of image conversion, image conversion applies the conversion of mathematic function variable domain essentially, processing image function with any mathematical model is the conversion to original image generally. The paper also briefly introduces the basic continuous Fourier transformation and discrete Fourier transformation, etc.Except for taking better image visual effect, the medical image processing has three destinations:display the complete and literal image structure has obtained clearly with proper image contrast; restrain or remove the picture noise effectively; display the important or key feature of the original image from some kinds of angle.In fact, the medical image can gained with some kinds of technical methods,which quality are good generally, but there are still some images of normal or poor quality. The poor quality images mainly lack of proper image contrast, or the image noise is too big, or the image minute structure can not be displayed clearly and visual effect. The medical CT images of poor quality are always in this way.In view of the destination of above image manipulation application, which emphasis should be adjust image contrast, remove image noise and image sharpening which can display important image feature.Enclosing the emphasis of medical application, the thesis firstly introduces the principle of image grayscale conversion and the application in CT image, so as to adjust the image grayscale contrast. Except for linear transform and non-linear transform (derives the logarithm transform mathematical model and exponential transform mathematical model of non-linear transform), the essay also proposes a histogram extending method to adjust image grayscale contrast and introduces the CT image application. This thesis mainly discusses about the aspects of linear transform and non-linear exponential transformation and logarithms transform, and programming with VB code, make linear transformation and non-linear exponential transformation and also logarithms transformation to the CT images of poor quality, so as to gain the CT image of good quality with better contrast. And also explains the CT window technique which belongs to linear transformation especially, and also slightly deploys discussion on its medical application.Removing picture noise is another application key point of medical CT image application. The thesis introduces the average smoothing processing of neighbor average smoothing processing and weighting average smoothing to reduce picture noise, and also smoothing processing application from general digital image to serious noise CT image. According to the result of programming and operating procedure on smoothing processing to general digital image images and CT images with big noise, the smoothing processing can reduce the influence of image noise to some degree and make the image soft (vague), but it can not reduce the noise completely and reduces the image contrast by contrary. Median filtering is a good method to remove image noise. By arranging the pixels from small to big (or big to small) and replacing the pixel value as median, the median filtering can produce median filtering image. It can reduce the noise of normal digital image and CT image effectively and improve the image contrast theoretically.Another point is the application of image sharpening. The edge of medical images area always full of biological information, but the edges are not so clear and even vague. So it is necessary to enhance the edges and make the image features outstanding. And sharpening is the technique to reinforce image edges, It can improve visual discriminating analysis effect; It consists of space domain processing and frequency domain processing. Space domain processing normally applies difference algorithm and frequency domain processing applies enhancing or protruding the highpass filtering of high-frequency amount method.From the angle of medical application, this essay introduces the edge enhancement, shadow filter, profile filtering and un-sharpening masking method. Edge enhancement can make the edges more clear; shadow filter can display the edges and shadows around by three-dimensional embossment; profile filtering can display the edges details; the basic algorithm of un-sharpening masking method is: g(x,y)=f(x,y)+c[f(x,y)-f(x,y)], f/(x, y) is original image,f(x,y) is man-made method, (like smoothing) vague; f(x,y) is the generated image; C is constant, g(x,y) is the sharpened image by anti-sharpen. The reason for sharpening enhancement is the high frequency components of vague image f(x,y) is greatly reduced, so when f(x,y) minus f(x, y), it means the f(x,y)-f(x,y) in bracket reduced greatly the low frequency components off(x,y), and the high frequency components is remained. Therefore, when c[f(x,y)-f(x,y)] is added to f(x,y), the high frequency component and edge is enhanced. When this sharpening technique is applied for processing images, the generated images can display the feature of original image from every angle, so as to make the doctor read more information. The paper mainly introduces disposal application of medical CT image.This thesis omits the theory system of medical image processing application, and discusses the processing application of medical (CT) images. This image processing technique is extremely useful in medical practical application, but it's a pity that few people apply this technique in practical medical. Take the CT image for example:in fact most of doctors only read the original CT image, and the original image has less information than the processed image(It can display the image of the original image from various angle,various flank, various characteristic). So it is quite necessary to spread and apply this technique in practical medical.Identification and processing of artifact is not referred in this paper.