Method for interactively drawing heterogeneous translucent object
A translucent, non-homogeneous technology, applied in the processing of 3D images, details involving image processing hardware, image data processing, etc., can solve practical application difficulties, can only represent relatively simple geometric shapes, and cannot achieve real-time rendering or Interactive rendering and other issues
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0082] Embodiment 1. The inventor has experimented with the entire user interaction and drawing pipeline: first, the input object model and non-homogeneous translucent material parameters are obtained, and the program draws a high-precision translucent object image in real time; then, the user performs the cutting operation , the object is cut into two halves, one half is displayed to the user and the cutting surface is reconstructed; third, the user edits the material on the cutting surface (such as using a brush), and the program interactively gives the effect of material editing; Fourth, the user chooses to reconstruct the original object model, and the program gives the drawing result of the original object model after material editing. All calculations are speeded up to interactive level.
[0083] The inventors also tested with different shapes, different translucent materials, and different grid densities (up to 1 million tetrahedrons), and found that the method of the p...
Embodiment 2
[0085] Embodiment 2. The inventor compared the drawing result of the invention with the drawing result of solving the diffusion equation by the finite element method. The two are almost identical in visual effect, and are very similar to the reference picture obtained by the Monte Carlo photon mapping method. The inventor calculated the point-by-point errors of the two results, and found that the maximum relative error of the two results was no more than 2%. The inventor tested the rendering efficiency of the two methods on a 4-core CPU at the same time, and found that the linear system construction process of the finite element method takes 24 seconds, while the linear system construction process of the present invention only takes 1 second. Note that if the linear system is constructed using the finite element method, the linear system can also be efficiently solved by the parallel multi-grid acceleration technology of the present invention.
Embodiment 3
[0086] Embodiment 3. The inventor compared the drawing results of the invention with the drawing results of solving the diffusion equation by the finite difference method. It was found that the finite-difference method requires a very high-precision tetrahedral mesh of the object due to its precision limitation. Using the finite difference method to calculate and draw the model after cutting will produce serious inaccurate drawing phenomenon. Although these inaccurate drawing can be reduced by careful mesh optimization, the time cost of optimization is such that the drawing efficiency will not reach interactive level. Even with the original uncut mesh, the finite difference method may still produce some visible artifacts, since the tetrahedral mesh entered by the user may not be highly optimized. According to the inventor's experimental results, the present invention is close to the finite difference method in terms of efficiency (implemented on the same GPU at the same time)...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 