Particle Video: Long-Range Motion Estimation Using Point Trajectories

By Peter Sand, Seth J. Teller. Computer Vision and Pattern Recognition, 2006

513 citations recorded by the source corpus.

Subfields: Computer Graphics and Computer-Aided Design, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/f43fe15d04f220b6d537d35a8412903acb4700f6

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Graphics and Computer-Aided Design

In Computer Vision and Pattern Recognition

Citations in the corpus

Object Segmentation by Long Term Analysis of Point Trajectories (2010, methodology): “An alterative tracker that will probably work as well with our technique is the one from [4], though the missing treatment of large displacements might be a problem in some sequences.”
Pixel-Level Hand Detection in Ego-centric Videos (2013, methodology): “[22] using the KLT tracker [26, 13] in place of the particle video tracker [21].”
Dense Point Trajectories by GPU-Accelerated Large Displacement Optical Flow (2010, methodology): “We also propose a method for dense point tracking by building upon the fast implementation of large displacement optical flow.”
A Fast Approximation of the Bilateral Filter Using a Signal Processing Approach (2006, background): “, 2004], optical flow and motion estimation [Xiao et al., 2006; Sand and Teller, 2006], and video processing [Bennett and McMillan, 2005; Winnemöller et al.”
Beyond pixels: exploring new representations and applications for motion analysis (2009, methodology): “Techniques such as iterative reweighted least squared (IRLS) [22, 20], pyramid-based coarse-to-fine search [14, 22] and occlusion/outlier detection [95] were intensively used for optimizing these nonlinear objective functions.”
CoTracker: It is Better to Track Together (2023, background): “PIPs [15] revisits the classic Particle Video [34] problem by introducing a model for point tracking through occlusions.”
CoTracker3: Simpler and Better Point Tracking by Pseudo-Labeling Real Videos (2025, unknown): “The task of tracking any point (TAP) was introduced by PIPs [12], who revisited the classic Particle Video [29] method and proposed using deep learning for point tracking.”
VGGT: Visual Geometry Grounded Transformer (2025, unknown): “Tracking-Any-Point was ﬁrst introduced in Particle Video [ 59] and revived by PIPs [32] during the deep learning era, aiming to track points of interest across video sequences, including dynamic motion.”
Motion Detail Preserving Optical Flow Estimation (2012, unknown): “Recently, the nonlocal smoothing strategy [33], [38] demonstrated the potential to handle displacement discontinuities and occlusion, which is tightly linked to explicit refinement of the flow field using image filtering [28], [40].”
A Lagrangian Particle Dynamics Approach for Crowd Flow Segmentation and Stability Analysis (2007, background): “Another related work is by Sand et. al., [ 17 ] which proposes a particle based framework for motion computation.”
Tracking-Learning-Detection (2012, background): “Various representations of the object are used in practice, for example, points [2], [3], [4], articulated models [5], [6], [7], contours [8], [9], [10], [11], or optical flow [12], [13], [14].”
A Quantitative Analysis of Current Practices in Optical Flow Estimation and the Principles Behind Them (2013, background): “3 and σe = 20; this is the same as in [53].”
Secrets of optical flow estimation and their principles (2010, background): “(22) in [29], I(i, j) is the color vector in the Lab space, and σ1 = 7, σ2 = 7.”
Ieee Transactions on Pattern Analysis and Machine Intelligence Segmentation of Moving Objects by Long Term Video Analysis (2014, unknown): “Other trackers are proposed in [9], [54].”
Abnormal crowd behavior detection using social force model (2009, background): “Furthermore, the particle advection captures the continuity of the crowd flow which neither optical flow nor any instantaneous measure could capture [24] [2].”
Background Subtraction for Freely Moving Cameras (2009, methodology): “The trajectories in these sequences were created using the particle video tracker of Sand and Teller [24].”
Human computing and machine understanding of human behavior: a survey (2006, unknown): “[50], [56]).”