Image Processing For A Part Of Virtual Chinese Human Data

"Chinese Digitized Virtual Human" is an advanced research field. Developed with combination of medicine and information science, it aims to visualize the structure and function of human body. Some groups of human data have been acquired in the country, and wide ranges of analysis and application on them have been in progress. In this study, some general methods for image progressing are summarized, while some new ones used in the manifold VCH data progressing, and several 3D models are obtained based on VCH data. In the end, we discuss advantage and disadvantage of these methods, and give some advices for VCH image progressing. Objective:1. Planar image progressing with the purpose of better image-shown and more convenience for storage and transfer. Image pre-processing for 3-D reconstruction will be done. At the same time, improve traditional methods for VCH image progressing.2. 3-D reconstruction based on VCH data with the methods of surface reconstruction and volume reconstruction. According to the results from different technique, find out the problems on data visualization with the traditional ways, and give some advices to the work of visualizing VCH data.Methods:1. Many methods are used to finish VCH planar image processing, including image geometry transform, gray-scale enhancement format transition and compression , image registration and segmentation.2. Surface reconstruction and volume reconstruction, the two methods are used to rebuild VCH data, which includes three groups of images: CT, MRI and true-color section images.Results:1. Image enhancement based on its histogram has been done, and convert DICOM format to JPG format in CT and MRI data progressing with a computer program self-made. Expected results are obtained in the work of image registration and image segmentation with two methods respectively.2. Finish 3-D reconstruction from VCH data with Matlab and other software, and models are accepted.Conclusions:1. Traditional methods and techniques of image progressing are helpful in the field of VCH, mainly CT and MRI data. New methods should be developed to process the most important data-true color section images, especially the methods for image registration and image segmentation with true color images. Besides, it is an urgent and hard problem to value the segmentation results in VCH data progressing.2. There are two methods used for medical image reconstruction, surface reconstruction and volume reconstruction, and each has some mature algorithms fitting the study for 3-D visualization with VCH data. Considering its most advantage, volume reconstruction should be the preferred method, but it carries complex calculation. At the same time, it is a difficult problem to solve that VCH data is too large. Surface reconstruction needs less calculation and can be carried fast with good-shown, but it loses internal information, so it is recommended at proper occasions.