A combined post-?ltering method to improve accuracy of variational optical ?ow estimation

摘要

We present a novel combined post-?ltering (CPF) method to improve the accuracy of optical ?ow estimation. Its attractive advantages are that outliers reduction is attained while discontinuities are well preserved, and occlusions are partially handled. Major contributions are the following: First, the structure tensor (ST) based edge detection is introduced to extract ?ow edges. Moreover, we improve the detection performance by extending the traditional 2D spatial edge detector into spatial-scale 3D space, and also using a gradient bilateral ?lter (GBF) to replace the linear Gaussian ?lter to construct a multi-scale nonlinear ST. GBF is useful to preserve discontinuity but it is computationally expensive. A hybrid GBF and Gaussian?lter (HGBGF) approach is proposed by means of a spatial-scale gradient signal- to-noise ratio (SNR) measure to solve the low ef?ciency issue. Additionally, a piecewise occlusion detection method is used to extract occlusions. Second, we apply a CPF method, which uses a weighted median ?lter (WMF), a bilateral ?lter (BF) and a fast median ?lter (MF), to post-smooth the detected edges and occlusions, and the other ?at regions of the ?ow ?eld, respectively. Benchmark tests on both synthetic and real sequences demonstrate the effectiveness of our method.