mpLBP: A point-based representation for surface pattern description

The Local Binary Pattern (LBP) is a very popular pattern descriptor for images that iswidely used to classify repeated pixel arrangements in a query image. Several extensionsof the LBP to surfaces exist, for both geometric and colorimetric patterns. Thesemethods mainly differ on the way they code the neighborhood of a point, balancingthe quality of the neighborhood approximation with the computational complexity. Forinstance, using mesh topological neighborhoods as a surrogate for the LBP pixel neighborhoodsimplifies the computation, but this approach is sensitive to irregular vertexdistributions and/or might require an accurate surface re-sampling. On the contrary,building an adaptive neighborhood representation based on geodesic disks is accurateand insensitive to surface bendings but it considerably increases the computational complexity.Our idea is to adopt the kd-tree structure to directly store a surface describedby a set of points and to build the LBP directly on the point cloud, without consideringany support mesh. Following the LBP paradigm, we define a local descriptor ateach point that is further used to define a global statistical Mean Point LBP (mpLBP)descriptor. When used to compare shapes, this descriptor reaches state of the art performances,while keeping a low computational cost. Experiments on benchmarks anddatasets from real world objects are provided altogether with the analysis of the algorithmparameters, property and descriptor robustness.