Online multiple classifier boosting for object tracking

Kim, Taekyun; Woodley, Thomas; Stenger, Björn; Cipolla, Roberto

doi:10.1109/cvprw.2010.5543889

Cited by 30 publications

(15 citation statements)

References 15 publications

(35 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This will increase both the efficiency and robustness of the classification. Although there are previous works that simultaneously learn to align and classify (Babenko et al, 2008;Kim et al, 2010), in our work we follow a different path. We find clusters of face poses after face detection and use them in training.…”

Section: Introductionmentioning

confidence: 99%

Robust gender recognition by exploiting facial attributes dependencies

Bekios-Calfa

Buenaposada

Baumela

2014

Pattern Recognition Letters

View full text Add to dashboard Cite

Estimating human face gender from images is a problem that has been extensively studied because of its relevant applications. Recent works report significant drops in performance for state-of-the-art gender classifiers when evaluated "in the wild," i.e. with uncontrolled demography and environmental conditions. We hypothesize that this is caused by the existence of dependencies among facial demographic attributes that have not been considered when building the classifier. In the paper we study the dependencies among gender, age and pose facial attributes. By considering the relation between gender and pose attributes we also avoid the use of computationally expensive and fragile face alignment procedures. In the experiments we confirm the existence of dependencies among gender, age and pose facial attributes and prove that we can improve the performance and robustness of gender classifiers by exploiting these dependencies.

show abstract

Section: Introductionmentioning

confidence: 99%

Robust gender recognition by exploiting facial attributes dependencies

Bekios-Calfa

Buenaposada

Baumela

2014

Pattern Recognition Letters

View full text Add to dashboard Cite

show abstract

“…Take Wang's tracker [8] for example, a spatial-color mixture of Gaussians appearance model was presented to encode both spatial layout and color information. For Kim's tracker [9], GMM was employed to weight the influence of strong classifier which judges whether each input bounding domain belongs to object region. The usage of GMM in this paper is essentially different from others, including object model, motion model, classifying mechanism, and online learning.…”

Section: Discussion About Gmmmentioning

confidence: 99%

On-Line Rigid Object Tracking via Discriminative Feature Classification

Miao

Shi

Meng³

et al. 2016

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYThis paper proposes an on-line rigid object tracking framework via discriminative object appearance modeling and learning. Strong classifiers are combined with 2D scale-rotation invariant local features to treat tracking as a keypoint matching problem. For on-line boosting, we correspond a Gaussian mixture model (GMM) to each weak classifier and propose a GMM-based classifying mechanism. Meanwhile, selforganizing theory is applied to perform automatic clustering for sequential updating. Benefiting from the invariance of the SURF feature and the proposed on-line classifying technique, we can easily find reliable matching pairs and thus perform accurate and stable tracking. Experiments show that the proposed method achieves better performance than previously reported trackers.

show abstract

“…Although they propose the scheme to cluster the object classes that may share the classification knowledge, we use it to automatically cluster target samples and learn cluster-specific appearance models, so that every target samples can be well discriminated from the nontarget samples by at least one of the multiple appearance models. As mentioned earlier, this scheme is more practical than that in [10] in solving the multi-modality problem they tackle, in the sense that the scheme automatically decides the number of clusters based on the amount of appearance variations formerly shown by the target. In contrast to [10], none of the off-line setups to construct the cluster priors is required.…”

Section: Learning P(y T |X T )mentioning

confidence: 99%

“…Then each feature is injected into one of the multiple trackers as an appearance model. In [10], the MCBoost is used to jointly learn the target sample clusters and cluster-specific classifiers as appearance models. In [14,13], the sparse coding scheme is used to extract the multiple templates from given training samples.…”

Section: Introductionmentioning

confidence: 99%

“…2) By slightly modifying the learning scheme in [17], jointly clustering target samples and learning cluster-specific appearance models are performed in a fully unsupervised manner. Different from the previous work [10] related to this topic, which requires the off-line setups to construct the cluster priors, the learning scheme adopted in the present paper automatically determines the number of clusters based on the amount of variations formerly shown by the target. This property makes the proposed tracker more practical in tracking generic objects with different amounts of appearance variations.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Robust visual tracking using autoregressive hidden Markov Model

Park

Kwon

Lee

2012

2012 IEEE Conference on Computer Vision and Pattern Recognition

View full text Add to dashboard Cite

Recent studies on visual tracking have shown significant improvement in accuracy by handling the appearance variations of the target object. Whereas most studies present schemes to extract the time-invariant characteristics of the target and adaptively update the appearance model, the present paper concentrates on modeling the probabilistic dependency between sequential target appearances ( Fig. 1-(a)). To actualize this interest, a new Bayesian tracking framework is formulated under the autoregressive Hidden Markov Model (AR-HMM), where the probabilistic dependency between sequential target appearances is implied. During the learning phase at each time step, the proposed tracker separates formerly seen target samples into several clusters based on their visual similarity, and learns clusterspecific classifiers as multiple appearance models, each of which represents a certain type of the target appearance. Then the dependency between these appearance models is learned. During the searching phase, the target state is estimated by inferring the most probable appearance model under the consideration of its dependency on formerly utilized appearance models. The proposed method is tested on 12 challenging video sequences containing targets with abrupt appearance variations, and demonstrates that it outperforms current state-of-the-art methods in accuracy.

show abstract

Online multiple classifier boosting for object tracking

Abstract: Abstract

Cited by 30 publications

References 15 publications

Robust gender recognition by exploiting facial attributes dependencies

Robust gender recognition by exploiting facial attributes dependencies

On-Line Rigid Object Tracking via Discriminative Feature Classification

Robust visual tracking using autoregressive hidden Markov Model

Contact Info

Product

Resources

About