Mixture models for optical flow computation

By Allan D. Jepson, Michael J. Black. 1993

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

On the Binding Problem in Artificial Neural Networks (2020, methodology): “Alternatively, the decomposition can be based on other features such as optical flow (Jepson and Black, 1993) or the coefficients of a wavelet transform (GuerreroColón et al., 2008).”
Efficient sparse-to-dense optical flow estimation using a learned basis and layers (2015, methodology): “Layered models [13, 23, 42, 45, 46, 49] provide another approach.”
Layered image motion with explicit occlusions, temporal consistency, and depth ordering (2010, methodology): “Wang and Adelson [25] cle arly motivate layered models of image sequences, while Jepson and Black [11] formalize the proble m using probabilistic mixture models.”
Layered RGBD scene flow estimation (2015, background): “Numerous papers have been published on RGB layered models, such as [4, 12, 22, 23, 46].”
Layered segmentation and optical flow estimation over time (2012, background): “Most layered approaches assume a parametric motion for each layer [11, 13, 16, 31] which is too restrictive to capture the motion of natural scenes.”
A Fully-Connected Layered Model of Foreground and Background Flow (2013, methodology): “Layered models [8, 12, 30] are promising for motion analysis, particularly for handling occlusion and capturing temporally consistent scene structure [27].”
Competitive Collaboration: Joint Unsupervised Learning of Depth, Camera Motion, Optical Flow and Motion Segmentation (2018, methodology): “In fact, the use of expectation-maximization (EM) in computer vision began with the optical flow problem and was used to segment the scene into “layers” [15] and was then widely applied to other vision problems.”
Robust online appearance models for visual tracking (2001, methodology): “the EM-algorithm for optical flow described in [16].”
Video Segmentation via Object Flow (2016, methodology): “Layered models [7, 15, 16, 34, 37, 46] jointly optimize for segmentation and optical flow.”
A unified mixture framework for motion segmentation: incorporating spatial coherence and estimating the number of models (1996, methodology): “Existing EM motion algorithms [8, 1] update gk(r) by calculating at”
Optical Flow with Semantic Segmentation and Localized Layers (2016, methodology): “To model complex scene motions and to deal well with motion boundaries, we adopt a layered approach [4, 11, 18, 21, 24, 49, 54, 55, 56].”
Estimating Optical Flow in Segmented Images Using Variable-Order Parametric Models With Local Deformations (1996, background): “Instead, this paper makes the simple hypothesis that image regions of piecewise smooth brightness are likely to correspond to surfaces in the world.”
Learning flexible sprites in video layers (2001, methodology): “A 3-dimensional scene is decomposed into a set of 2- dimensional objects in layers, hugely simplifying the geometry [1] [2] [3] [4].”
A Database and Evaluation Methodology for Optical Flow (2007, methodology): “Most methods have focused on the use of parametric models that estimate motion in layers [34,77].”
Human motion analysis: a review (1997, methodology): “It is an extension to motion analysis of rigid objects using the expectation-maximization (EM) algorithm [ 21 ].”
Optical Flow Estimation (2006, methodology): “resentation of a scene in which different cardboard surfaces correspond to different layers, and they are assumed to be able to move independently [31, 49].”