Motion Detail Preserving Optical Flow Estimation

By Li Xu, Jiaya Jia, Yasuyuki Matsushita. 2012

428 citations recorded by the source corpus.

How researchers see this paper

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

Citations in the corpus

Flow Fields: Dense Correspondence Fields for Highly Accurate Large Displacement Optical Flow Estimation (2015, background): “Since then, several works followed the idea of using sparse descriptors [34, 32, 20, 28, 26].”
Motion estimation via hierarchical block matching and graph cut (2015, methodology): “We first detect the occluded pixels based on the observation that, multiple pixels mapping to the same point in the target frame using forward warping are possibly occluded by each other [12].”
Unifying Flow, Stereo and Depth Estimation (2022, background): “Optical flow has been traditionally tackled with variational approaches [2], [7], [39], [40], [41], [42], where it is typically solved as an energy minimization problem that consists of a brightness constancy term and a regularization term.”
PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume (2017, methodology): “To address this issue, Brox and Malik [10] embed feature matching into the variational framework, which is further improved by follow-up methods [50, 56].”
Context-Aware Synthesis for Video Frame Interpolation (2018, methodology): “Bidirectional flow-guided blending MDP-Flow2 [50]”
Video Segmentation via Object Flow (2016, methodology): “To compute accurate optical flow fields, it is common to segment images or extract edges to preserve motion details around object boundaries [3, 43, 44, 47].”
Efficient sparse-to-dense optical flow estimation using a learned basis and layers (2015, methodology): “[53] use sparse SIFT features to generate additional translational flow hypotheses.”
Super SloMo: High Quality Estimation of Multiple Intermediate Frames for Video Interpolation (2017, methodology): “In addition to SepConv, we also compare our model with other three top-performing models on the Middlebury dataset2, where the interpolation algorithm [1] is coupled with different optical flow methods including MDP-Flow2 [37], PMMST [39], and DeepFlow [34].”
An Overview to Visual Odometry and Visual SLAM: Applications to Mobile Robotics (2015, unknown): “A number of algorithms to solve the optical ﬂow problem using motion constraint equations have been proposed (see [115] for a list of current approaches).”
Layered RGBD scene flow estimation (2015, background): “Some notable variants include image-dependent, non-local motion priors to recover sharp motion boundaries [38, 47, 51], fusion moves to reach stronger local minima [24], embedding feature matching to track small, fast-moving objects [10, 45, 49], exploiting the epipolar constraint for egocentric sce”
Unsupervised Deep Learning for Optical Flow Estimation (2017, methodology): “Xu et. al (Xu, Jia, and Matsushita 2012) further integrates more advanced matching methods such as SIFT-flow (Liu, Yuen, and Torralba 2011) and PatchMatch (Barnes et al. 2010) to increase the accuracy of flow.”
Phase-based frame interpolation for video (2015, methodology): “A translation of this function can be described by modifying the phase, e.g. by subtracting π/4 in our example.”
Layered segmentation and optical flow estimation over time (2012, methodology): “This can partly be remedied by explicit occlusion detection and post processing [37].”
Video Frame Interpolation via Adaptive Convolution (2017, methodology): “MDP-Flow2 [56], which currently produces the lowest interpolation error according to the Middlebury benchmark, the method from Brox et al.”
Occlusion Aware Unsupervised Learning of Optical Flow (2017, result): “On MPI-Sintel, our method has a comparable result with previous non-neural-network based methods [33, 57].”
Fast Optical Flow Using Dense Inverse Search (2016, methodology): “More details in § 3.1 and recent endeavors lead to signiﬁcant progress in handling large displacements [45, 13, 11, 31, 49, 28, 36, 41, 2, 51, 53, 54, 55, 18].”
EpicFlow: Edge-preserving interpolation of correspondences for optical flow (2015, background): “State-of-the-art approaches [9, 38] use descriptor matching between adjacent frames together with the inte-”