Voxeles

Voxeles are units of volumetric space in a three-dimensional grid, analogous to pixels in two dimensions. In a voxel-based representation, the three-dimensional scene is discretized into small cube-shaped elements, each carrying properties such as color, density, or material type. This approach records volume data directly, rather than deriving geometry from surfaces.

The term voxel comes from a combination of “volume” and “pixel.” In practice, voxeles are used to store scalar or vector fields across a bounded region, enabling direct manipulation of volume data. Data can be organized in dense voxel grids, where every position within a bounding box has a value, or in sparse structures that store only non-empty regions to save memory. Common sparse representations include sparse voxel octrees, voxel hashing, and compressed formats that support efficient access, editing, and rendering at varying levels of detail.

Rendering techniques for voxeles include ray casting and ray marching, which trace rays through the voxel grid to accumulate color and illumination, as well as splatting and voxel-based global illumination methods. Real-time visualization often relies on hierarchical data structures (such as octrees) and voxel cone tracing or related approaches to balance quality and performance.

Voxeles offer advantages such as straightforward boolean operations, easy modeling of volumetric phenomena, and natural handling of internal structures and semi-transparent materials. Their drawbacks include high memory requirements at fine resolutions and challenges in producing smooth surfaces without additional surface extraction or interpolation. Applications span medical imaging (CT and MRI data), scientific visualization, geology and materials science, as well as games and simulations that employ voxel-based environments and effects.