https://hgpu.org/?p=4213
Implementations of Parallel Computation of Euclidean Distance Map in Multicore Processors and GPUs