Improvement And Application Of X-ray Medical Image Algorithm Retinex Theory Base

Roentgen discovered x-rays in1895,and at a very short period of time, X ray has been widely used, which is also used in medical imaging. Applications of rays have been promoting the development of medical science and clinical medicine. Especially these recent20years, with the combination of computers, myriads of methods of treatments have been created which lead us into an era of tomography, digital, and3D simulation and reconstruction. As a consequence, all kinds of image processing methods as fundamentals have been advanced dramatically.Compared with common images, X-ray medical image has its particularity with the most significant character for its wider image dynamic range, richer details, and lower contrast. Therefore, striding to newer image improving means should be taken into consideration except those conventional digital image processing methods such as histogram equalization, spatial filter, and frequency domain filter and so on. Retinex theory is one of those methods with profits of increasing image dynamic range, enhancing image brightness, and improving visual effects of the image which is appropriate to the complementary of medical images characterized in lower contrast rate, poorer resolution, and worse visual effect. According to the characteristics of medical image, it put forward an improved algorithm LRA (Logsig cumulative Reintex Algorithm) based on Retinex theory for X-Ray medical image.Thus, this article focuses on these items as bellow:Firstly, traditional image processing methods are analyze and their characteristics are studied;Secondly, constructed noise model and applied the last denoising method to reduce the noise of X-ray medical image;Thirdly, analyzed Retinex theory and programmed classic Retinex algorithms along the development of the theory. It can get better brightness when used the neural network logarithmic S form LogSig transmission function compared with logarithm function and then proposed personal LRA Retinex Algorithm based on those. Via comparison between the new-proposed method and classic ones, we find out drawbacks when applying the traditional Retinex algorithms on X-ray medical images and point out the newer method’s profits of improving the image quality and advantages contrasted to former ones.