Kernel-Based Object Tracking

By Dorin Comaniciu, Visvanathan Ramesh, Peter Meer. 2003

5,099 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

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

Temporal Fusion Transformers for Interpretable Multi-horizon Time Series Forecasting (2019, methodology): “To compare similarities between attention weight vectors, we use the distance metric proposed by [9]: where ρ ( p , q ) = (cid:205) j √ p j q j is the Bhattacharya coefficient [22] measuring the overlap between discrete distributions – with p j , q j being elements of probability vectors p , q…”
Newtonian Image Understanding: Unfolding the Dynamics of Objects in Static Images (2015, background): “Tracking: Note that our approach is quite different from tracking [17, 11, 10] since tracking methods are not destined for single image reasoning.”
Robust Visual Tracking via Structured Multi-Task Sparse Learning (2012, methodology): “The mean shift tracker (Comaniciu et al. 2003) is a popular mode-finding method, which successfully copes with camera motion, partial occlusions, clutter, and target scale variations.”
Coupled Object Detection and Tracking from Static Cameras and Moving Vehicles (2008, methodology): “The second subproblem is typically addressed by classic tracking approaches such as Extended Kalman Filters (EKF) [15], particle filtering [21], or Mean-Shift tracking [6], which rely on a Markov assumption and carry the associated danger of drifting away from the correct target.”
Multiple kernel tracking with SSD (2004, background): “The basic idea of kernel-based estimation can now be easily extended to recover richer models of target motion following the line of development generally used for template tracking [4, 1, 9].”
Monocular Model-Based 3D Tracking of Rigid Objects: A Survey (2005, background): “Furthermore, an excellent reference text for this topic already exists [54].”
A survey on activity recognition and behavior understanding in video surveillance (2013, methodology): “[30] used a Kalman filter framework based on the color histogram in conjunction with a kernel-based target localization method for tracking human beings in the subway.”
Context tracker: Exploring supporters and distracters in unconstrained environments (2011, methodology): “However, Meanshift and color information are vulnerable in a cluttered background and the fusion of multiple components, i.e. the auxiliary objects in this method, is ad-hoc.”
Face tracking and recognition with visual constraints in real-world videos (2008, background): “Numerous approaches to target modeling have attempted to tackle these issues using view-based appearance models [4], contour models [9], 3D models [11], mixture models [10], and kernel representations [6, 8], among others.”
Online Object Tracking: A Benchmark (2013, methodology): “In addition to template, many other visual features have been adopted in tracking algorithms, such as color histograms [16], histograms of oriented gradients (HOG) [17, 52], covariance region descriptor [53, 46, 56] and Haar-like features [54, 22].”
Segmentation and Tracking of Multiple Humans in Crowded Environments (2008, methodology): “By using the Epanechikov profile [6], LðuÞ will be increased, with the new location”
A Novel Performance Evaluation Methodology for Single-Target Trackers (2015, neutral): “filter, a background model as in [66] and information coming”
Tracking Using CamShift Algorithm and Multiple Quantized Feature Spaces (2004, methodology): “Kernel based object tracking (including adaptive scale and background-weighted histogram extensions) is described by Comaniciu et al .”
Robust object tracking via sparsity-based collaborative model (2012, methodology): “Meanwhile, the representation schemes can be based on holistic templates [6] or local histograms [1, 28].”
Robust Fragments-based Tracking using the Integral Histogram (2006, background): “Second, the geometric relations between the template patches allow us to take into account the spatial distribution of the pixel intensities - information which is lost in traditional histogram-based algorithms.”
Tracking-Learning-Detection (2012, background): “Template tracking can be either realized as static [15]”
Fast multiple object tracking via a hierarchical particle filter (2005, methodology): “In contrast, color-based trackers are faster and more robust, where the color histogram is typically used to model the targets to combat the partial occlusion, and non-rigidity [27, 37, 3, 8].”