Snappable Meshes Performance Dataset

<strong>Overview</strong> This dataset contains performance and navigation benchmarks obtained by generating eight maps with the snappable meshes algorithm multiple times. Data in columns is organized as follows: <code>run</code> - Number of generation run. <code>genset</code> - Generated map (maps "(a)" to "(h)" from the Benchmark scene. <code>navset</code>- Number of navigation points used to generate the navigation metrics. <code>tg</code> - Duration of the map generation process, in milliseconds. <code>tv</code> - Duration of the map validation process, in milliseconds. <code>c</code> - Average percentage of valid connections between navigation points. <code>ar</code> - Relative area of the largest fully-connected (i.e., fully-navigable) region. <code>nclu</code> - Number of isolated regions, i.e., of regions which are not connected to any another. <code>genseed</code> - Seed used for the generation process. <code>navseed</code> - Seed used for determining the validation metrics. A number of results presented in the research paper "Procedural Generation of 3D Maps with Snappable Meshes" are obtained from this dataset. <strong>Reproducibility of results</strong> The results presented in the research paper "Procedural Generation of 3D Maps with Snappable Meshes", namely in Figure 8 and Figure 10, can be reproduced with the Jupyter notebook included with this dataset (file <code>analysis.ipynb</code>). <strong>Licenses</strong> The dataset is made available under a CC-BY 4.0 license (see <code>LICENSE_DATA.txt</code>). The code in the Jupyter Notebook is made available under the MIT license (see <code>LICENSE_CODE.txt</code>).