![the sims 3 cc tiles and mosaic sculpture the sims 3 cc tiles and mosaic sculpture](https://www.thesimsresource.com/scaled/1902/w-600h-450-1902725.jpg)
In this example we did not use any texture information, assigning random colors to the tiles. The bear mosaic sculpture in Figure 10 illustrates the variation in size of tiles on the 3D surface according to the approximated local curvature of the mesh. The Voronoi tiles spread over more than one polygon from the original mesh but nevertheless there are no visible discontinuities on the Voronoi cells. In Figure 9, we can see that the planar approximation of the Voronoi computation handles well the covering of the mesh. First we show a result on a mesh with the same overall curvature, a sphere.
![the sims 3 cc tiles and mosaic sculpture the sims 3 cc tiles and mosaic sculpture](https://www.thesimsresource.com/scaled/1479/w-920h-574-1479521.jpg)
Our renderings emphasize just the overall distribution of tiles and did not use any particular realistic rendering material model, except in Figure 11 where a wood-like material was used. 25 GB RAM are around 10 sec for 15000 tiles, mostly due to computation of the Voronoi cells. In Table 1 we list the parameters used: number of tiles, mapping of tile sizes to curvatures, and amount of grout. In this section we illustrate a few results of our technique. After the Voronoi computation, as illustrated in Figure 8, the tiles are properly aligned with the edges from the texture, and the final configuration still follows the opus palladium mosaic style. Particles will be pushed away from the edges, and due to repulsive forces between them, they will be forced to align near edges. This force can be multiplied by a scaling factor w that better suits the particular number of tiles being used. We then map this vector into the coordinate system of the base polygon and add it to the resulting repulsive force due to the neighboring particles.
![the sims 3 cc tiles and mosaic sculpture the sims 3 cc tiles and mosaic sculpture](https://www.thesimsresource.com/scaled/2497/w-867h-650-2497945.jpg)
Subtracting 177 from each, we obtain directly the force vector in the coordinate system of the texture. For each particle, we query the blue and green components from the computed texture. Once we have the texture with the information about the gradient, we are able to use it in our particle relaxation method. Figure 7(b) shows the resulting texture computed as explained above with d r = 0. Negative values will be in the range and positive values in the range. The equations below compute the gradient in the blue B ( i, j ) and green G ( i, j ) components respectively, of the new texture: Zero values of the gradient will be mapped to 177.