Font Size: a A A

Research On 3D Visual Reconstruction Algorithm Based On ITK And VTK

Posted on:2019-03-22Degree:MasterType:Thesis
Country:ChinaCandidate:D Y LiFull Text:PDF
GTID:2334330545490169Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
China is the country with the highest incidence of liver diseases in the world.Visualization and 3D reconstruction technology as a supplementary medical treatment is one of the key technologies for liver diagnosis and treatment.The main process is to automatically segment the liver and lesions first,and then visualize the segmentation results,thus providing more intuitive image guidance for the doctor's clinical diagnosis,and improve the diagnostic accuracy.Based on this background,this article has done research on the automatic segmentation and visual reconstruction efficiency of liver in the process of three-dimensional reconstruction.The main contents and innovations of this study are as follows:(1)An automated segmentation algorithm for abdominal liver and lesions is proposed.Cascading two full convolutional neural networks,using 3DIRCAD to expose the data set as training data,obtained a feature model.The segmented structure of the obtained liver and lesions was combined with conditional random fields,and finally a more accurate segmentation result was obtained.(2)A fast algorithm for time optimization of surface-drawing mobile cubes is proposed.During the drawing process,the ergodic conditions of the voxels are constrained to reduce the extraction of equipotential surfaces such as empty voxels.Using the segmented 3DIRCAD dataset,a three-dimensional reconstruction image of the liver is obtained.Compared with the liver image drawn directly using the mobile cube algorithm,the reconstruction time is shortened and the reconstruction effect is better.The experimental results show that the automatic segmentation accuracy of the liver in the data set 3DIRCAD can reach 94%with the combination of the cascaded full convolutional neural network and the conditional random field;and the reconstruction time of the surface rendering time optimization algorithm is proposed to have a 10%reduction in the reconstruction time;According to the reconstruction effect,accurate segmentation can effectively remove the effects of other organs on the reconstruction of liver and lesions.
Keywords/Search Tags:Medical image, Full convolutional neural network, Visualization and 3D reconstruction
PDF Full Text Request
Related items