Segmentation And Intelligent Recognition Of Nerve Fibers

The disease of peripheral nerve injury is very common in our daily life, however, to fully recover under the present stage is very difficult. With the improvement of people’s living standard, people are expecting to cure the disease. In the current hospital treatment of this disease, the key lies in how to make the same properties of nerve bundle stitching together, but in this process has a relatively difficult problems:the nerve bundle are of the same nature, how to determine if they are of the same nature. Doctors find a way to solve this problem. First photograph to injury of peripheral nerve, get the corresponding microscopic image, identify the same properties of artificial observation of microscopic image of neural function, but the process is often very complex and huge effort. In order to solve this problem, some experts have thought that use computer to do the repetitive work, through the computer to do all kinds of microscopic image processing peripheral nerve, including image segmentation, recognition, finally identified the same nature of neural function beam help doctors perform operations. Before this, the study of CT image is more than the study of microscopic image, almost no research related the microscopic image of peripheral nerve, and the microscopic image 3 d reconstruction of the peripheral nerve has a lot to help doctors cure disease of peripheral nerve injury, therefore, need to in-depth study.In this paper, adopting digital morphology combined with classic image segmentation algorithm on peripheral nerve bundle of nerve fibers in the segmentation, extract the corresponding feature for recognition. And on the basis of a large number of simulation, obtained some good results, using C++language for peripheral nerve system to realize 3 d reconstruction. In this paper the research content is as follows:1. Because pure Karnorvsky-Roots method dyeing microscopic image peripheral nerve defect has many defects, put forward three another redyeing image methods. These methods can be a good way to avoid using single Karnorvsky-Roots dyeing defects, and gives a method to these three redyeing image steps, finally introduced the steps for peripheral nerve microscopic image.2. Segmentation of nerve fibers. Using digital technology related to morphology to improve the region growing algorithm, and using it to split nerve fibers. This algorithm firstly takes into consideration the problem of a large number of nerve fibers adhesion, and after experiments, the results confirm the validity of the algorithm.3. Feature extraction of nerve fibers. Using the doctor’s experience judgment and the nature of the specific of microscopic image, extract nerve fiber features. Take advantage of these features in three peripheral nerves of redyeing experiments on microscopic image, get the corresponding results of nerve fibers.4. Peripheral nerve system design and implementation of 3 d reconstruction. According to the object-oriented program design idea, design the function frame of the whole system, and show the results after the difficult problem to solve.This paper is to solve the three-dimensional reconstruction of neural function in peripheral nerve bundle of nerve fibers in the segmentation and recognition. This is the last step in 3 d reconstruction of peripheral nerve, and it is also the most difficult step. In this paper, the improved algorithm is put all kinds of nerve fibers separated, in other words, is to portray the outline of nerve fibers. After separated by extracting the relevant identification of nerve fibers, finally through the computation nerve bundle of nerve fibers ratio, complete the 3 d reconstruction of peripheral nerve.