Background Subtraction for Freely Moving Cameras

By Yaser Sheikh, Omar Javed, Takeo Kanade. IEEE International Conference on Computer Vision, 2009

366 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/b322b3ecb1b8d2264641388ebcf1d64fb1aa0cb2

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

Actionness Estimation Using Hybrid Fully Convolutional Networks (2016, background): “There have been several works on action proposal generation [25, 13, 46], but there is no standard evaluation protocol to evaluate these different proposal generation algorithms.”
Pixel-Level Hand Detection in Ego-centric Videos (2013, methodology): “Baseline: Feature trajectory projection [22] using KLT”
Background Subtraction in Real Applications: Challenges, Current Models and Future Directions (2019, methodology): “In addition, background subtraction can be also used in applications in which cameras are slowly moving [322][98][343].”
Segmenting Salient Objects from Images and Videos (2010, positive): “These circumstances have been addressed by introducing adaptive background models and methods to eliminate camera movements [11, 12], but both of these are difficult problems and technically demanding.”
Ego2Hands: A Dataset for Egocentric Two-hand Segmentation and Detection (2020, background): “Before the revolution of deep convolutional networks in the field of computer vision, [30, 15, 34] proposed motion-based approaches for binary foreground segmentation with the assumption that the motion pattern is different for the foreground and background.”
Action recognition in videos acquired by a moving camera using motion decomposition of Lagrangian particle trajectories (2011, background): “Since the camera motion subspace is approximately spanned by three basis [21, 25], the camera motion component can be estimated by Ac = US V ′ , where U and V are obtained by singular value decomposition [U, S, V ] = SV D(A), and S is equal to S except that all the singular values other than the mos”
Video Object Segmentation through Spatially Accurate and Temporally Dense Extraction of Primary Object Regions (2013, methodology): “A subset of these methods employs motion grouping ([19, 18, 4]) for object segmentation.”
Improving RGB-D SLAM in dynamic environments: A motion removal approach (2017, methodology): “In [19], the method estimated a set of trajectories by tracking sparse salient points across video frames.”
Moving Object Detection by Detecting Contiguous Outliers in the Low-Rank Representation (2011, background): “An alternative approach for motion segmentation tries to se gment the objects by analyzing point trajectories [9], [22], [23], [24].”
Traditional and recent approaches in background modeling for foreground detection: An overview (2014, background): “Background subtraction can be used in applications in which cameras are slowly moving [227][87][245][313].”
Actionness Ranking with Lattice Conditional Ordinal Random Fields (2014, background): “[32], which does not seek to differentiate between general motion and action at all.”
Video Processing From Electro-Optical Sensors for Object Detection and Tracking in a Maritime Environment: A Survey (2016, background): “Number of videos 4 32 Total number of frames 1196 16254 Number of frames in a video 299 [206, 995]”
Analysis of Hand Segmentation in the Wild (2018, methodology): “, using global hand template [29, 24, 28], and (3) motion-based detection; using ego-motion of the camera and assuming hands (foreground) and the background have different motion [27, 9, 8].”
Figure-ground segmentation improves handled object recognition in egocentric video (2010, background): “[20] builds a sparse representation for background feature trajectories and subtracts it from dynamic scenes.”
Model Recommendation with Virtual Probes for Egocentric Hand Detection (2013, methodology): “As in many segmentation techniques, we formulate the task of hand region contour segmentation as an energy minimization problem [1] over super-pixel regions [13, 15, 5].”