Representing moving images with layers

By John Y. A. Wang, Edward H. Adelson. 1994

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

Attribute2Image: Conditional Image Generation from Visual Attributes (2015, methodology): “1D representations of images have been studied in the context of moving or still object segmentation [38,23,39,41,11].”
LR-GAN: Layered Recursive Generative Adversarial Networks for Image Generation (2016, background): “Representing images in layered structure is even straightforward for video with moving objects (Darrell & Pentland, 1991; Wang & Adelson, 1994; Kannan et al., 2005).”
Piecewise Rigid Scene Flow (2013, background): “In the context of stereo disparity and optical flow, explicit modeling of discontinuities by means of segmentation or layer-based formulations has a long history [23] and has recently gained renewed attention: Bleyer et al.”
Object tracking: A survey (2006, methodology): “Motion-based segmentation [Vidal and Ma 2004; Black and Anandan 1996; Wang and Adelson 1994] and multibody factorization [Costeira and Kanade 1998; Gear 1998] methods have been used to learn models for multiple objects moving in a scene.”
Ieee Transactions on Pattern Analysis and Machine Intelligence Segmentation of Moving Objects by Long Term Video Analysis (2014, unknown): “While early approaches estimate the optical flow and the segmentation independently [69], [56], optical flow estimation and segmentation were later considered as a joint optimization problem [25], [3], [15], [63].”
Background Subtraction for Freely Moving Cameras (2009, methodology): “Layer-based methods [33, 28, 15, 35] model the scene as piece-wise planar scenes, and cluster segments based on some measure of motion coherency.”
Figure-ground segmentation improves handled object recognition in egocentric video (2010, background): “[4]) at a computational cost, and motion layers can be extracted from the flow (classical works including [25, 1]).”
Seeded region growing: an extensive and comparative study (2005, methodology): “…in estimating a dense motion field and then partition the scene only based on the obtained motion information, where the adjacent video components are merged to form the meaningful video objects if they obey the same Hough or affine transformation motion model (Adiv, 1985; Wang and Adelson, 1994).”
Epitomic analysis of appearance and shape (2003, unknown): “We bound the log likelihood of the data using an approximate posterior of the form q = q(s1)q(s2)q(m), log p(x|es, em) > B = = ∫ s1,s2,m q(s1)q(s2)q(m) log p(x, s1, s2,m|es, em) q(s1)q(s2)q(m) = ∫ s1,s2,m q(s1)q(s2)q(m) log p(x|s1, s2,m) + ∫ s1 q(s1) log p(s1|es) + ∫ s2 q(s2) log p(s2|es) + ∫ m…”
Greedy Learning of Multiple Objects in Images Using Robust Statistics and Factorial Learning (2004, background): “Some early work on this topic is by Wang and Adelson (1994) , and an example of more recent work is that of Tao, Sawhney, and Kumar (2000).”
AutoFlow: Learning a Better Training Set for Optical Flow (2021, methodology): “We take a layered approach [49, 44] to rendering image pairs and their optical flow, as shown in Fig.”
Video matting of complex scenes (2002, methodology): “1997] or automated [Wang and Adelson 1994] means, the remaining background fragments can be assembled using perspective image mappings to form a composite mosaic [Szeliski and Shum 1997], which can then be reprojected into each original frame to form a dynamic clean plate.”
Synthesizing Images of Humans in Unseen Poses (2018, background): “Moving scenes can be understood as a composition of layers [25].”
Articulated pose estimation with tiny synthetic videos (2015, background): “[37, 35], where each pixel is augmented by its depth layer.”
Learning generative models of mid-level structure in natural images (2012, background): “Layered representations that represent an image in terms of several overlapping objects which are arranged according to a depth ordering have been suggested at least 20 years ago (Adelson, 1991; Wang and Adelson, 1994) and have, since then been extensively used and formulated as probabilistic, gener”
Coded exposure photography: motion deblurring using fluttered shutter (2006, background): “Techniques such as image inpainting hallucinate the possible pixel values [Wang and Adelson 1994][Black and Anandan 1996], but encoded blur allows us to encompass a completely observed system.”
Optical flow modeling and computation: A survey (2015, background): “Similarly, layered approaches [9,223,231,234,257,277] aim at decomposing the observed scene into overlapping layers and introduce information about depth ordering of regions.”