期刊论文

Tu, Zhigang

英文

Pattern Recognition

0031-3203

2017

65

11-25

10.1016/j.patcog.2016.10.027

Our new objective model is a function of variables (u, v, I1, I2). To simultaneously solve all the unknowns, an alternating minimization framework [28] can be used. Since the optical flow terms (i.e. the first and second terms of Eq. (14)) depend on the restored images, and the image restoration terms (i.e. the second and third terms of Eq. (14)) depend on the computed optical flow, we split the objective function into two coupled modules: EO(u,v,I)=λ∫ΩΨS(∥▿u∥2+∥▿v∥2)dxdy+∫ΩΨD((I2(x+u,y+v)−I1(x,y

本校为其他机构

The most popular optical flow algorithms rely on optimizing the energy function that integrates a data term and a smoothness term. In contrast to this traditional framework, we derive a new objective function that couples optical flow estimation and image restoration. Our method is inspired by the recent successes of edge-aware constraints (EAC) in preserving edges in general gradient domain image filtering. By incorporating an EAC image fidelity term (IFT) in the conventional variational model, the new energy function can simultaneously estima...