Real-time tracking of non-rigid objects using mean shift

By Dorin Comaniciu, Visvanathan Ramesh, Peter Meer. Proceedings IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2000 (Cat. No.PR00662), 2000

3,395 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/2cfa006b33084abe8160b001f9a24944cda25d05

How researchers see this paper

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

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Tracking across multiple cameras with disjoint views (2003, neutral): “This is the modified Bhattacharyya coefficient [4].”
Learning Adaptive Discriminative Correlation Filters via Temporal Consistency Preserving Spatial Feature Selection for Robust Visual Object Tracking (2018, methodology): “Subsequently, mean-shift [21] was proposed for visual object tracking with iterative histogram matching.”
Multisensor-Based Human Detection and Tracking for Mobile Service Robots (2009, methodology): “Moreover, the integration of an additional visual tracker, such as the histogram-based solution of Comaniciu et al. [ 29 ], could help in keeping the tracks apart even when the faces are not detectable.”
Pedestrian detection and tracking with night vision (2005, methodology): “Pedestrian-tracking techniques that were proposed recently include matching of middle lines of candidates by Bayesian classification [6], spatial‐temporal trajectory patterns and Kalman filter prediction [29], combination of stereo parallax, color consistency and face patterns [10], meanshift method”
Visual tracking with online Multiple Instance Learning (2009, neutral): “Although many tracking methods employ static appearance models that are either defined manually or trained using the first frame [16] [8] [1], these methods tend to have difficulties tracking objects that exhibit significant appearance changes.”
Template Matching with Deformable Diversity Similarity (2016, methodology): “For example in [3, 21] color Histogram Matching is used (for tracking).”
Local Trinary Patterns for human action recognition (2009, background): “This is a common practice in object recognition, e.g., [ 3 ].”
Visual tracking decomposition (2010, methodology): “We compared the proposed algorithm (VTD) with four different tracking methods: standard MCMC (MC) based on [9][16]; Mean Shift (MS) [2] based on the implemented function in OpenCV; On-line Appearance Learning (OAL) in [17]; and Multiple Instance Learning (MIL) in [1].”
Recent advances and trends in visual tracking: A review (2011, background): “The state of this art has advanced significantly in the past 30 years [1–8].”
Region Covariance: A Fast Descriptor for Detection and Classification (2006, methodology): “Following [6], histograms were widely used for nonrigid object tracking.”
Color-Based Probabilistic Tracking (2002, methodology): “(“MeanShift” in [5]), and modified later by Chen et al.”
Robust Object Tracking with Online Multiple Instance Learning (2011, positive): “Although many tracking methods employ static appearance models that are either defined manually or trained using only the first frame [2], [4], [6], [7], [8], [9], these methods are often unable to cope with significant appearance changes.”
Coupled minimum-cost flow cell tracking for high-throughput quantitative analysis (2011, background): “Kalman filters [2], particle filters [3], and mean shift tracking [4] are typical examples of stochastic filters.”
UvA-DARE (Digital Academic Repository) Siamese Instance Search for Tracking (2016, methodology): “MST [6] relies on probabilistic matching.”
The Visual Object Tracking VOT2016 Challenge Results (2016, methodology): “PKLTF consists of two phases: The first one uses the Kanade Lukas Tomasi approach (KLT) [43] to choose the object features (using colour and motion coherence), while the second phase is based on mean shift gradient descent [44] to place the bounding box into the position of the object.”
Matching Pixels Using Co-occurrence Statistics (2018, methodology): “When tracking a deformable object it might be better to represent the template as a histogram and use an appropriate similarity measure between histograms [2].”
Mean Shift: A Robust Approach Toward Feature Space Analysis (2002, background): “For e xampl , in [14] by defining the distance between the distributions of the model and a candidate of t he target, non-rigid objects were tracked in an image sequence under severe distortions.”