Random ensemble metrics for object recognition

Kozakaya, Tatsuo; Ito, Satoshi; Kubota, Susumu

doi:10.1109/iccv.2011.6126466

Cited by 20 publications

(13 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although one normal vector from support vector machine (SVM) that amounts to projecting the SPM onto 1-D subspace may be enough for the binary classification problem, it is far less enough for visual tracking due to the inherent challenges. Kozakaya et al [15] presented an ensemble metric learning scheme which learns multiple discriminative projection vectors obtained from linear SVM using randomly subsampled training data. In our work, we find a suitable metric matrix in the feature space of sparse codes via online metric learning rather than resorting to linear SVMs to adaptively capture appearance variations.…”

Section: Related Workmentioning

confidence: 99%

Metric Learning Based Structural Appearance Model for Robust Visual Tracking

Yang

et al. 2014

IEEE Trans. Circuits Syst. Video Technol.

View full text Add to dashboard Cite

Appearance modeling is a key issue for the success of a visual tracker. Sparse representation based appearance modeling has received an increasing amount of interest in recent years. However, most of existing work utilizes reconstruction errors to compute the observation likelihood under the generative framework, which may give poor performance, especially for significant appearance variations. In this paper, we advocate an approach to visual tracking that seeks an appropriate metric in the feature space of sparse codes and propose a metric learning based structural appearance model for more accurate matching of different appearances. This structural representation is acquired by performing multiscale max pooling on the weighted local sparse codes of image patches. An online multiple instance metric learning algorithm is proposed that learns a discriminative and adaptive metric, thereby better distinguishing the visual object of interest from the background. The multiple instance setting is able to alleviate the drift problem potentially caused by misaligned training examples. Tracking is then carried out within a Bayesian inference framework, in which the learned metric and the structure object representation are used to construct the observation model. Comprehensive experiments on challenging image sequences demonstrate qualitatively and quantitatively that the proposed algorithm outperforms the state-of-the-art methods.

show abstract

Section: Related Workmentioning

confidence: 99%

Metric Learning Based Structural Appearance Model for Robust Visual Tracking

Yang

et al. 2014

IEEE Trans. Circuits Syst. Video Technol.

View full text Add to dashboard Cite

show abstract

“…In this paper, we learn from [10] random ensemble metrics (REMetric) method, combined with the ensemble learning based on support vector machine. Support Vector Machine (SVM) is firstly proposed by Cortes and Vapnik in 1995, which seeks to find the maximum separation hyper-plane of two categories, therefore over-fitting phenomenon is rarely seen, and it show unique advantages in pattern recognition of high dimensional feature space.…”

Section: Introductionmentioning

confidence: 99%

A Novel Tracking Method Based on Ensemble Metric Learning

Huo

2014

2014 Tenth International Conference on Computational Intelligence and Security

View full text Add to dashboard Cite

In recent years, object tracking is often formulated as detection tasks. Although many methods of machine learning and statistical learning can effectively achieve discriminate target model, but they usually require a large amount of training samples for satisfactory precision. Combining advantages of ensemble learning and the support vector machine, this paper proposes a simple tracking method based on metric learning. With a small number of training data sampled from the sequence of images during the tracking, we are learning an adaptive metric matrix that tends to maximum the distance between samples of different classes. In the process of tracking, metric matrix is learned and updated constantly to achieve good discrimination and adaptability. Experimental results show that our method has better tracking stability and accuracy.

show abstract

“…To this end, there have been increasing efforts made to learn a distance metric in recent years [6,10,37,39,41,50]. Metric learning methods can be categorized into unsupervised [50], semisupervised [5] and supervised ones [1,11,25,27,34,53], according to the availability of the labels of training samples. Metric learning has been proved to successfully improve the clustering and recognition performance in information retrieval [29,30], bioinformatics [47] and computer vision tasks [6,10,12,16,39,37].…”

Section: Introductionmentioning

confidence: 99%