Research On Optical Computed Tomography Image Reconstruction Based On BP Neural Network

In recent years, Optical Computed Tomography as a new imaging technology has attachted much attention in medical field. This technology which uses Near Infrared Ray Source gains superiority to any other existing imaging methods. In order to solve the problem that it is difficult to put this method into clinical application, a new image reconstruction of Optical Computed Tomoraphy based on BP neural network is propsed in this paper. Solve the inverse problem as an optimization approach. Then, the fast image reconstruction of the map of tissue optical parameters in brain or galactophore has been realized.Firstly, the diffusion theory and the definition of forward problem and inverse problem in optical imaging are introduced. At the same time, FEMLAB software is employed to construct the forward model, and the simulation data is provided to BP neural network as training set. Then an improved BP neural network which employed to solve the inverse problem is constructed based on the forward model.Through the difference between the measured and the predicted data, the optimization method of optical properties is made. An improved BP neural network based on Levenberg-Marquardt optimization is used to solve the inverse problem. Interactive validation method was used to test the output of BP neural network. Through times of training and parameters adjusting, the network has been obtained. At last, Phantom experiments were done to validate the reconstructed results. Then the reconstructed optical properties were put into FEMLAB simulation, the map of tissue optical distributions can be obtained. From the comparison between theory simulation and Phatom experiments, the method of BP neural network has been confirmed.In this paper, the optical Computed Tomography image reconstruction based on BP neural network has been confimed. The fast image reconstruction of optical parameters is realized, and the reconstruct results are satisfied. This method lay a solid foundation for the future research, and has important meaing. This original image reconstruction method is also provide a new idear to other imaging method, and is valued for the image reconstruction technology of medical imaging.