Learning flexible sprites in video layers

By Nebojsa Jojic, Brendan J. Frey. 2001

309 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

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

In Media Technology

Citations in the corpus

Variational Inference: A Review for Statisticians (2016, background): “Some of the earliest examples included inferring nonlinear image manifolds (Bishop and Winn 2000) and finding layers of images in videos (Jojic and Frey 2001).”
Neural Expectation Maximization (2017, background): “Early work in unsupervised video segmentation [17] used generalized Expectation Maximization (EM) to infer how to split frames of moving sprites.”
Greedy Learning of Multiple Objects in Images Using Robust Statistics and Factorial Learning (2004, methodology): “7 These data are used in Jojic and Frey (2001). We thank N.”
Optical Flow with Semantic Segmentation and Localized Layers (2016, methodology): “Several methods decompose scenes into layers corresponding to objects [22, 24, 28, 53, 60].”
Video (language) modeling: a baseline for generative models of natural videos (2014, background): “Perhaps with the only exception given by the layered model proposed by Jojic & Frey (2001), all the above mentioned models have been demonstrated on either small image patches or small synthetic datasets (Sutskever et al., 2009; Pachitariu & Sahani, 2012; Michalski et al., 2014).”
Weakly-Supervised Action Localization With Background Modeling (2019, background): “Notable exceptions in the literature include probabilistic latent-variable models that build generative models of both foreground and background [16].”
Tracking Multiple Occluding People by Localizing on Multiple Scene Planes (2009, methodology): “For a detailed review of the state of the art in tracking research, the reader is referred to the recent survey by Yilmaz et al. [61].”
Beyond pixels: exploring new representations and applications for motion analysis (2009, background): “It is possible, in principle, to perform segmentation using motion alone [123, 53], but reliable segmentation requires the use of additional features.”
Motion magnification (2005, methodology): “It is possible, in principle, to perform segmentation using motion alone [Wang and Adelson 1994; Jojic and Frey 2001], but reliable segmentation requires the use of additional features.”
Bilayer Segmentation of Live Video (2006, unknown): “frames Figure 1: An example”
LOCUS: learning object classes with unsupervised segmentation (2005, methodology): “Following [12], we discretise the space of transformations and restrict it to all whole-pixel translations at a small fixed range of scales.”
Layered segmentation and optical flow estimation over time (2012, methodology): “Our depth ordering formulation is similar to flexible sprites [13] and their extensions [16], but we use more flexible motion models.”
Layered object detection for multi-class segmentation (2010, background): “Such approaches are typically explored in the video domain and include examples such as layered motion models [25], video sprites [10], and layered pictorial structures [12].”
A Fully-Connected Layered Model of Foreground and Background Flow (2013, background): “Jojic and Frey [13] extract segmented regions and reason about their depth order, but focus on simple translational motions and do not provide a segmentation of the scene.”
Robust online appearance models for visual tracking (2001, background): “One can improve the robustness of such trackers by representing the variability of each pixel in the template [12], [18].”