A Benchmark for the Comparison of 3-D Motion Segmentation Algorithms

By Roberto Tron, René Vidal. 2007 IEEE Conference on Computer Vision and Pattern Recognition, 2007

751 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/22effe373836d4feae2ce3e8abec0058438bebb9

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

A Benchmark Dataset and Evaluation Methodology for Video Object Segmentation (2016, background): “Other datasets exist, but they are mostly provided to support specific findings and thus are either limited in terms of total number of frames, [8,21,25,47], or do not exhibit a sufficient variety in terms of content [1, 4, 5, 15, 17, 20, 41, 46].”
Moving Object Detection by Detecting Contiguous Outliers in the Low-Rank Representation (2011, background): “When a new frame is arrived, it is projected onto the subspace spanned by the principle components, and the residues indicate the presence of new objects.”
A closed form solution to robust subspace estimation and clustering (2011, methodology): “We apply our framework to the Hopkins155 motion segmentation database [18], which is available online at http://www.”
Fast Approximate Energy Minimization with Label Costs (2010, methodology): “21 Rigid motion estimation examples (Vidal’s data set [56]) comparing different algorithms minimizing energy (?): α++, α+, α-SM, α-BM.”
Low rank subspace clustering (LRSC) (2014, methodology): “We use the Hopkins155 motion segmentation database (Tron and Vidal, 2007) to evaluate the performance of LRSC against that of other algorithms.”
Motion Segmentation in the Presence of Outlying, Incomplete, or Corrupted Trajectories (2010, methodology): “Fig. 4. Example image frames from three motion sequences from the Hopkins155 database [ 24 ].”
Sparse Subspace Clustering: Algorithm, Theory, and Applications (2013, methodology): “94 percent, reported in [66] and [35], to 4.”
Latent Low-Rank Representation for subspace segmentation and feature extraction (2011, methodology): “To evaluate the effectiveness of LatLRR, we test it on Hopkins155 [25], which is a standard database for motion segmentation.”
Robust Recovery of Subspace Structures by Low-Rank Representation (2010, methodology): “While several methods have achieved an error rate below 3% on Hopkins155, subspace segmentation problem is till far from solved.”
Learning Motion Patterns in Videos (2016, methodology): “We also evaluate on sequences from Berkeley (BMS-26) [5, 36] and Freiburg-Berkeley (FBMS) [26] motion segmentation datasets.”
Multiframe Motion Segmentation with Missing Data Using PowerFactorization and GPCA (2008, positive): “Our recent work ( Tron and Vidal 2007 ) shows that the method of Yan and Pollefeys (2005) can give improved performance for sequences with two and three motions.”
Visual SLAM and Structure from Motion in Dynamic Environments (2018, methodology): “s: Evaluated on synthetic data, r: Evaluated on real data, h: Evaluated on Hopkins 155 dataset [159].”
Object Segmentation by Long Term Analysis of Point Trajectories (2010, methodology): “While the Hopkins 155 dataset [8] has clearly boosted research in multi-body factorization, it is much too specialized for these types of methods, and particularly the checkerboard sequences do not correspond to natural scenes.”
Smooth Representation Clustering (2014, methodology): “Hopkins155 [25] is a motion segmentation dataset, consisting of 155 video sequences with extracted feature points and their tracks across frames.”
Robust Subspace Segmentation by Low-Rank Representation (2010, methodology): “In this section, we evaluate LRR on the Hopkins155 motion database (Tron & Vidal, 2007).”
Energy-Based Geometric Multi-model Fitting (2012, methodology): “The general geometric framework for (13) is to detect independent motions as hyperplanes, which is comparable to those for GPCA and RANSAC in Tron and Vidal (2007).”
Sparse Subspace Clustering: Algorithm, Theory, and Applications (2012, methodology): “94 percent, reported in [66] and [35], to 4.”
Video Processing From Electro-Optical Sensors for Object Detection and Tracking in a Maritime Environment: A Survey (2016, background): “Optical flow [19], [86], [91], [149], [201]−[204] Motion maps learnt for segmented objects or just features in image frames Can deal with occlusion, can use multiple features simultaneously Computation intensive, many dense motion layers for complex motion Feature tracking and clustering [201], [205”