Object Segmentation by Long Term Analysis of Point Trajectories

By Thomas Brox, Jitendra Malik. 2010

889 citations recorded by the source corpus.

Subfields: 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 Computer Vision and Pattern Recognition

Citations in the corpus

FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks (2016, methodology): “Motion Segmentation: We report results using [19, 14] on the training set of FBMS-59 [27, 19] with a density of 4 pixels.”
YouTube-VOS: A Large-Scale Video Object Segmentation Benchmark (2018, methodology): “In comparison to previous datasets [16,21,22,23,24,25], DAVIS has both higher-quality of video resolutions and annotations.”
Unsupervised Object Discovery and Tracking in Video Collections (2015, methodology): “To exploit this information, we build on long-term point tracks [3] and propose a motion coherence score for motion-based foreground conﬁdence.”
One-Shot Video Object Segmentation (2016, methodology): “has to compute temporal matches in the form of optical flow or dense trajectories [5], which can be an even harder problem.”
Unsupervised Detection and Tracking of Arbitrary Objects with Dependent Dirichlet Process Mixtures (2012, methodology): “Other attempts to perform unsupervised detection and tracking include methods for clustering short sequences of positions extracted by detecting the motion of objects [8, 9], which aim to return full-length distinct object tracks, and graph based methods that carry out a similar task using spectral ”
SfM-Net: Learning of Structure and Motion from Video (2017, methodology): “We show results of our approach on KITTI [12, 21], MoSeg [4], and RGB-D SLAM [27] benchmarks under different levels of supervision.”
Learning Unsupervised Video Object Segmentation Through Visual Attention (2019, background): “Early UVOS methods are typically based on handcrafted features and heuristics such as long-term point trajectory [54, 5, 17, 53, 9], motion boundary [56], objectness [43, 51, 89, 18, 59, 83, 40, 45], and saliency [13, 77, 76, 34, 27].”
Stereo Video Object Segmentation Using Stereoscopic Foreground Trajectories (2019, methodology): “There is a recent interest in exploiting trajectories for representing long-term motion information [4], [5], [9], [11], [13], [40], [44].”
A Continuous Occlusion Model for Road Scene Understanding (2016, methodology): “Some recent efforts include robust algebraic segmentation with hybrid perspective constraints [21] and spectral clustering with point track spatial affinities [1].”
Moving Object Detection by Detecting Contiguous Outliers in the Low-Rank Representation (2011, background): “TABLE III QUANTITATIVE EVALUATION USING THE SEQUENCES FROM THEBERKELEY MOTION SEGMENTATION DATASET[24].”
Video Pop-up: Monocular 3D Reconstruction of Dynamic Scenes (2014, methodology): “Evaluation of the Motion Segmentation Step: We evaluate the results of our object-level segmentation on the Berkeley Motion Segmentation Dataset using the tracks and evaluation tool proposed in [6].”
Spatio-Temporal Moving Object Proposals (2014, methodology): “We test our method on the two largest publicly available video segmentation benchmarks: VSB100 [14] and Moseg [8].”
Robust video segment proposals with painless occlusion handling (2015, neutral): “Some approaches utilize tracked feature points [5,6,14,29,34].”
Instance Embedding Transfer to Unsupervised Video Object Segmentation (2018, background): “Differences in optical flow can separate objects moving against a background [3].”
Consistent Video Saliency Using Local Gradient Flow Optimization and Global Refinement (2015, methodology): “Although videos from SegTrack [21] and FBMS [43], [44] databases span a large range of difﬁculties, the amount Fig.”
Learning object class detectors from weakly annotated video (2012, methodology): “3.2 we extract segments of coherent motion from each shot, us-ing the technique of Brox and Malik [8] (ﬁg.”
Video Segmentation by Tracking Many Figure-Ground Segments (2013, background): “Therefore, it is hard to strike correct temporal links without examining appearance information from a larger region [8], or temporal information over a longer period [7], both requiring extensive computations.”