抄録
A self-organizing map (SOM) neural network algorithm is applied to the epipolar line nearest-neighbor particle matching of the 3-D particle tracking velocimetry (PTV). Then, the same algorithm is again applied to the time-differential particle matching of the 3-D PTV for completing the 3-D particle tracking process. The principle of the self-organizing map method seems especially effective even if there are loss-of-pair particles between the two spatio-differential or time-differential frames because of the competitive learning strategy.
