Robust online appearance models for visual tracking

By Allan D. Jepson, David J. Fleet, Thomas F. El-Maraghi. Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001, 2001

1,253 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/4411f262853bf7f1eb8e2efe03eb0402f5e9ad2c

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

Object Tracking Benchmark (2015, methodology): “Effective update algorithms have also been proposed via online mixture model [29], online boosting [22], and incremental subspace update [47].”
Real-Time Compressive Tracking (2012, background): “Tracking algorithms can be generally categorized as either generative [1, 2, 6, 10, 9] or discriminative [3– 5, 7, 8] based on their appearance models.”
Superpixel tracking (2011, background): “On the other hand, numerous works have demonstrated that adaptive appearance models play a key role in achieving robust object tracking [9] [4] [13] [1] [11] [20].”
Contour-based object tracking with occlusion handling in video acquired using mobile cameras (2004, background): “Similarly, Jepson et al. [4] compute the affine motion of the object using a probabilistic appearance model that captures the stable object features, object shape, and deals with outliers.”
Object tracking using SIFT features and mean shift (2009, background): “Other tracking strategies can also be found as Multiple Hypothesis Tracking (e.g., [5,6]), kernel-based tracking (e.g., [7, 8 ]), and optical flow-based tracking (e.g., [9]).”
Robust Object Tracking with Online Multiple Instance Learning (2011, positive): “In this scenario, it has been shown that an adaptive appearance model, which evolves during the tracking process as the appearance of the object changes, is the key to good performance [10], [11], [12].”
Robust Visual Tracking using 1 Minimization (2009, methodology): “A mixture model of three components with an online EM algorithm is proposed to model the appearance variation during tracking [20].”
Kernel-Based Object Tracking (2003, background): “Learning of appearance models by employing a mixture of stable image structure, motion information, and an outlier process, was discussed in [41].”
Sparse coding based visual tracking: Review and experimental comparison (2013, background): “Jepson et al. [9] proposed a framework to learn an adaptive appearance model, which adapts to the changing appearance over time.”
Self-Paced Learning for Long-Term Tracking (2013, background): “Appearance models: Because object appearance is likely to change over time, many tracks update appearance models through color histogram tracking [10] and online adaption [11].”
Online Object Tracking: A Benchmark (2013, methodology): “Effective update algorithms have also been proposed via online mixture model [29], online boosting [22], and incremental subspace update [47].”
Learning Multi-domain Convolutional Neural Networks for Visual Tracking (2015, methodology): “Various generative target appearance modeling algorithms have been proposed including sparse representation [31, 43], density estimation [16, 22], and incremental subspace learning [32].”
Recent advances and trends in visual tracking: A review (2011, methodology): “Jepson et al. [70] developed an elaborate mixture model with an online EM algorithm to explicitly model the appearance changes during tracking.”
Visual tracking decomposition (2010, background): “The appearance changes cover geometric and photometric variations of an object such as occlusion, pose, or illumination changes [8, 17].”
Robust Visual Tracking via Structured Multi-Task Sparse Learning (2012, methodology): “In Jepson et al. (2003), a robust and adaptive appearance model is learned for motion-based tracking of natural objects.”
Incremental Learning for Visual Tracking (2004, methodology): “Experiments show t hat good tracking results can be obtained with this representation without resorting to wavelets as used in [10], and better performance can potentially be achieved using wavel et filters.”