Bladder Wall, Based On The Oct Image Denoising And Segmentation Algorithms,

The development of biomedicine tomography technology brings great benefits to medical diagnosis. The ultrasonic and nuclear magnetic resonance technologies are mature. With the emergence of new techniques, the medical science will definitely reach a new higher level. Optical Coherence Tomography (OCT) is a new tomography technique which uses the coherence feature of infrared ray and heterodyne detection form cross-section images for biological tissues. Compared with traditional tomography technology, OCT is high-resolution, high probing depth and it is widely applied to the diagnosis and treatment of ophthalmology and dermatology. However, the new technology can only applied to limited fields, the key of research focuses on whether it could exploit the advantage to the full.This paper firstly introduces the background of OCT technology and reviews the development of OCT. It defines the significance of this research combining the difficulties of OCT. Secondly, the paper analyses the difficulties and shortages in OCT image noise reduction and segmentation, and introduces the key point of the research. Then it raises a mixed filtering method based on wavelet transform and mean filter which is directed to speckle noise in the OCT images. This method analyses the reason of speckle noise and features of itself, then turns images into logarithmic domain, it changes speckle noise model and breaks down the image data by using wavelet transform. And then retains the low-frequency approximate image signal, using mean filter respectively to horizontal, vertical and diagonal high-frequency signals in three directions to reduce the noise. Finally, the denoised image will be obtained by the integration of every sub-signal.Then, the paper raises an improved region growing method based on texture analysis regarding segmenting OCT images of bladder wall, the purpose of which is to distinguish structure of bladder wall layer. This method utilizes mixed filtering method which was mentioned above. It conducts preprocessing for OCT original image, eliminating noises and strengthens boundary. Then it analyzes the local texture of the image and draws the texture features, based on which it defines the choice of seed point and formulation of growing criterion. Finally it gets the result following the above conditions.