EpicFlow: Edge-preserving interpolation of correspondences for optical flow

By Jérôme Revaud, Philippe Weinzaepfel, Zaïd Harchaoui, Cordelia Schmid. 2015

821 citations recorded by the source corpus.

We propose a novel approach for optical flow estimation, targeted at large displacements with significant occlusions. It consists of two steps: (i) dense matching by edge-preserving interpolation from a sparse set of matches; (ii) variational energy minimization initialized with the dense matches. The sparse-to-dense interpolation relies on an appropriate choice of the distance, namely an edge-aware geodesic distance. This distance is tailored to handle occlusions and motion boundaries - two common and difficult issues for optical flow computation. We also propose an approximation scheme for the geodesic distance to allow fast computation without loss of performance. Subsequent to the dense interpolation step, standard one-level variational energy minimization is carried out on the dense matches to obtain the final flow estimation. The proposed approach, called Edge-Preserving Interpolation of Correspondences (EpicFlow) is fast and robust to large displacements. It significantly outperforms the state of the art on MPI-Sintel and performs on par on Kitti and Middlebury.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Deep multi-scale video prediction beyond mean square error (2015, methodology): “To this end, we use the EpicFlow method of Revaud et al. (2015), and compute the different quality measures only in the areas where the optical ﬂow is higher than a ﬁxed threshold 2 .”
FlowNet: Learning Optical Flow with Convolutional Networks (2015, methodology): “The successive work of [30] termed EpicFlow has put even more emphasis on the quality of sparse matching as the matches from [35] are merely interpolated to dense flow fields while respecting image boundaries.”
Unsupervised Learning of Edges (2015, methodology): “As our goal is to test the quality of generated edge maps, we focus only on variants of EpicFlow [31], the highest performing method on Sintel as of CVPR 2015.”
Playing for Benchmarks (2017, methodology): “We benchmark the performance of four well-known optical ﬂow algorithms: LDOF [7], EpicFlow [51], Flow-Fields [3], and FlowNet [14].”
YouTube-VOS: A Large-Scale Video Object Segmentation Benchmark (2018, background): “Although some recent approaches [11,10,13] have been proposed to leverage temporal consistency, they depend on models pretrained on other tasks such as optical flow [17,18] or motion segmentation [19], to extract temporal information.”
Cross Pixel Optical Flow Similarity for Self-Supervised Learning (2018, positive): “Epic-Flow [51] + Deep-Matching [52] was used to compute opticalflow for these frames.”
Optical Flow Estimation Using a Spatial Pyramid Network (2016, methodology): “Deep matching methods [21, 32, 44] do not fully solve the problem, since they resort to classical methods to compute the final flow field.”
Learning Video Object Segmentation from Static Images (2016, methodology): “Given a video sequence, we compute the optical ﬂow using EpicFlow [37] with Flow Fields matches [2] and convolutional boundaries [27].”
Long-Term Image Boundary Extrapolation (2016, methodology): “In case of the optic flow baseline, the optic flow is calculated between the last and the second last input frames (the frames at time t - 1 and t) using the Epic flow method of [20].”
MEMC-Net: Motion Estimation and Motion Compensation Driven Neural Network for Video Interpolation and Enhancement (2018, methodology): “In addition to the above learning-based frame interpolation methods, we also use existing optical flow algorithms (SPyNet [11] and EpicFlow [10]) to directly”
Unsupervised Deep Learning for Optical Flow Estimation (2017, methodology): “2013) or EpicFlow (Revaud et al. 2015), and the supervised deep learning approach FlowNet (Fischer et al.”
Two-Stream Convolutional Networks for Dynamic Texture Synthesis (2017, background): “Since no large-scale flow dataset exists that captures natural imagery with groundtruth flow, we take an unlabeled video dataset and apply an existing flow estimator [35] to estimate optical flow for training, cf .”
Object Contour and Edge Detection with RefineContourNet (2019, methodology): “Resulting object contour maps are very useful for supporting and/or improving various computer vision applications, like semantic segmentation [5,31,35], object proposal [24] and object flow estimation [25,31].”
Unsupervised convolutional neural networks for motion estimation (2016, background): “Index Terms— Motion Estimation, Convolutional Neural Network, Unsupervised Training”
SegFlow: Joint Learning for Video Object Segmentation and Optical Flow (2017, methodology): “To better determine the correspondences between images, matching based optimization algorithms [43, 32] are developed, in which these methods usually require longer processing time.”
Deep Crisp Boundaries (2017, methodology): “We choose EpicFlow [30] as our optical flow method.”
Learning Correspondence From the Cycle-Consistency of Time (2019, background): “These issues have driven researchers to study methods for estimating long-range optical flow [56, 5, 55, 51, 52, 34].”
Spatio-Temporal Filter Adaptive Network for Video Deblurring (2019, methodology): “(b) is the optical flow from the adjacent input blurry frames (a) and (f) according to EpicFlow [33].”