Identification And Evaluation Of Specific Regions On Discrete Point Model

Fast and accurate initial treatment to trauma has a positive effect for the cure. Through theevaluation of the degree of trauma, patients for the treatment will be injected in accordance withthe reasonable dosage in infusion. Especially for large area burn patients, accurate rehydrationduring shock stage has a crucial role in maintaining the vital signs and wound healing, to a certainextent, it determines the patient survival after burns. And one of the main factors that decide thedegree of trauma is the wound area size. Therefore, quick and accurate evaluation of surfacedamaged area plays an vital role in clinical treatment.In this paper, the color3D point model of a human is obtained from construction by the3Ddiscrete and featured point data in physical space coordinate system and color information, whichare obtained from the3D scanner Kinect. First, the model is rendered as a three-dimensional effecton the window of the system, and the user is allowed to interact with the system to select somekey vertices corresponding the model on the screen. And the selected2D screen coordinate ismapped to the model in physical space coordinate system, in order to obtain the vertex of themodel. Then, with the specified key vertices, the region boundary which surrounds the specifiedregion is outlined quickly and accurately, using the geodesic curve or a convex curve. Thegeodesic curve and convex curve are aimed to basically match the skin region boundary. Aftergetting the more detailed region boundary, the growing seed is determined by using the method ofplane projection and locating the approximate center, and the specific region is rapidly extractedout. Finally, all vertices within the specified region are used for triangular meshing, to calculatethe specific area in size and in percentage, which are supported for the further evaluation andanalysis.This paper has put forward specifying the aimed area based on the discrete point model. Onone hand, it can overcome the defects from the traditional automatic segmentation, such that it isnot easy to control, even the segmentation results are not accurate. While interactive option cannot only explor the reasonal experienced judgement, but also can make the rapid lacation of thespecific region for the system. On the other hand, it can solve the previous problem that it is a rough result by the method of directly selecting triangles on the meshed model, and it can beapplied to process the model of sparse sampling. In addition, different from the existingassessment method for damaged area, a3D human body surface imaging technology is used inthis paper to effectively obtain3D physic space cooridates data and other information, such as theequipment Kinect which is small and light. It is fast, safe and accurate for the human modelingsystem based on Kinect to construct a three-dimensional model of the human body. The3D colormodel obtained from Kincet is used as the research object in this paper, it is fast and accurate fordata acquisition, and it is easy for later comparison or other purposes. In order to effectively dealwith large amounts of3D point, the advantage of KD tree is also fully used in this paper. KD treealgorithm is applied to calculate the vertices of the region border. And combining the KD-treeFLANN (fast neighborhood search algorithm) method, traditional methods of calculating thegeodesic curve and the convex curve on the model are improved, and so the performance of areaextraction.Finally, some experiments were done with the mannequin and the patched persons tocompare the extraction area based on the method in this paper with that of the existing methods.The results demonstrate the accuracy and effectiveness of this method. Therefore, in this paper,the study of area estimation based on3D color point model and interactive region growing regionextraction method provides a good support to the fast, convenient and accurate assessment of bodydamaged area.