This was a project for CIS 5660, Procedural Graphics. The goal of this project was to lego-ify a mesh so that it "blockifies" it - making it composed of a variety of lego blocks. Throughout this project, I used Houdini's Rubber Toy Test Geometry node to show the output of the blockified mesh.

The first step was to convert the given mesh to a volume of points. I did this by using a VDB node and a Convert VDB node to compute the volume of the mesh and then using a Points from Volume node to generate the points in 3D space from the mesh. Next, I set the scale of all the particles so that the particle separation (how far points are from each other) can be customized by the user. This impacts how "blocky" the end output is. Larger particle separation results in larger lego bricks and a "blockier" feel to the output.

Below are two outputs obtained from blockifying - the first one is a result of smaller particle separation between points and the second one uses a larger scale.

The lego output uses a variety of bricks: Block bricks (2x2, 2x1, 1x1), slope bricks (1x1) and grill bricks (2x1).

The logic for categorizing points is done inside an Attribute Wrangle node. I essentially create three groups, one for sloped bricks, grill bricks, and block bricks, and assign each point to one of the three groups.

Slope bricks are placed on particles where the normal is sufficiently dissimilar to the up vector. Grill bricks are placed on particles where there are no particles above the current one. This can be determined by using the pcfind function, and offsetting the current y position by a small amount as the location to search. Then, when pcfind returns a group of nearby points, we can check to see if they are points above the current particle. Lastly, we place block points on all other particles.

However, because the LEGO blocks used are different sizes and shapes, we use a greedy method to first place as many large blocks as possible and then fill the space with smaller blocks. This is done by iterating through each point and keeping track of which points remain unused by a brick after assigning bricks to particles as described above. Once we have finished our for loop, we will have split the original input group of points into "points occupied by LEGO bricks" and "points unoccupied by LEGO bricks". Since we cannot place any more of the current size of LEGO bricks onto our points, then we must occupy the remaining unused points with a smaller brick type. We can do this repeatedly with decreasing brick sizes until we have occupied all points with a brick.