Multi-Scale Generalized Plane Match for Optical Flow

Despite recent advances, estimating optical flow remains a challenging problem in the presence of illumination change, large occlusions or fast movement. In this paper, we propose a novel optical flow estimation framework which can provide accurate dense correspondence and occlusion localization through a multi-scale generalized plane matching approach. In our method, we regard the scene as a collection of planes at multiple scales, and for each such plane, compensate motion in consensus to improve match quality. We estimate the square patch plane distortion using a robust plane model detection method and iteratively apply a plane matching scheme within a multi-scale framework. During the flow estimation process, our enhanced plane matching method also clearly localizes the occluded regions. In experiments on MPI-Sintel datasets, our method robustly estimated optical flow from given noisy correspondences, and also revealed the occluded regions accurately. Compared to other state-of-the-art optical flow methods, our method shows accurate occlusion localization, comparable optical flow quality, and better thin object detection.

• Publication year

  2018

• Venue

  arXiv.org

• Publication date

  2018-04-11

• Fields of study

  Computer Science

• Identifiers
  arXiv 1804.03787
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Coarse-to-Fine PatchMatch for Dense Correspondence

  2018cited by this paper
• Supplementary material of : CNN-based Patch Matching for Optical Flow with Thresholded Hinge Embedding Loss

  2017cited by this paper
• Robust Interpolation of Correspondences for Large Displacement Optical Flow

  2017cited by this paper
• S2F: Slow-to-Fast Interpolator Flow

  2017cited by this paper
• Efficient Coarse-to-Fine Patch Match for Large Displacement Optical Flow

  2016cited by this paper
• A Variational Aggregation Framework for Patch-Based Optical Flow Estimation

  2016cited by this paper
• Full Flow: Optical Flow Estimation By Global Optimization over Regular Grids

  2016cited by this paper
• SPM-BP: Sped-Up PatchMatch Belief Propagation for Continuous MRFs

  2015cited by this paper
• DeepMatching: Hierarchical Deformable Dense Matching

  2015cited by this paper
• PatchBatch: A Batch Augmented Loss for Optical Flow

  2015cited by this paper
• EpicFlow: Edge-preserving interpolation of correspondences for optical flow

  2015cited by this paper
• Local Layering for Joint Motion Estimation and Occlusion Detection

  2014cited by this paper
• Statistical Atlases and Computational Models of the Heart - Imaging and Modelling Challenges

  2014cited by this paper
• Fast Edge-Preserving PatchMatch for Large Displacement Optical Flow

  2014cited by this paper
• Highly Overparameterized Optical Flow Using PatchMatch Belief Propagation

  2014cited by this paper
• PM-Huber: PatchMatch with Huber Regularization for Stereo Matching

  2013cited by this paper
• Large Displacement Optical Flow from Nearest Neighbor Fields

  2013cited by this paper
• A Naturalistic Open Source Movie for Optical Flow Evaluation

  2012cited by this paper
• PatchMatch Stereo - Stereo Matching with Slanted Support Windows

  2011cited by this paper
• PatchMatch: a randomized correspondence algorithm for structural image editing

  2009influential reference
• Occlusion-Aware Optical Flow Estimation

  2008cited by this paper
• Optimal Randomized RANSAC

  2008cited by this paper
• Bilateral Filtering-Based Optical Flow Estimation with Occlusion Detection

  2006cited by this paper

• No citing papers are available for this paper.