Abstract
The distance transform (DT) and its many variations are ubiquitous tools forimage processing and analysis. In many imaging scenarios, the images ofinterest are corrupted by noise. This has a strong negative impact on theaccuracy of the DT, which is highly sensitive to spurious noise points. In thisstudy, we consider images represented as discrete random sets and observestatistics of DT computed on such representations. We, thus, define astochastic distance transform (SDT), which has an adjustable robustness tonoise. Both a stochastic Monte Carlo method and a deterministic method forcomputing the SDT are proposed and compared. Through a series of empiricaltests, we demonstrate that the SDT is effective not only in improving theaccuracy of the computed distances in the presence of noise, but also inimproving the performance of template matching and watershed segmentation ofpartially overlapping objects, which are examples of typical applications whereDTs are utilized.