Large Margin Local Metric Learning

Bohné, Julien; Ying, Yiming; Gentric, Stéphane; Pontil, Massimiliano

doi:10.1007/978-3-319-10605-2_44

Cited by 30 publications

(18 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Figure 2 shows examples of an image affected by the three levels of noise we tested: none, medium and strong. We compare our method, Uncertainty-Aware Joint Bayesian (UA-JB), to three other methods: Joint Bayesian (JB) (Chen et al, 2012) to which our method is equivalent in the absence of noise, ITML (Davis et al, 2007) and LMLML (Bohné et al, 2014) in single metric mode. We start by reducing the dimensionality to 100 using UA-PPCA for UA-JB and standard PCA for the three others as prescribed by the authors.…”

Section: Mnistmentioning

confidence: 99%

“…For each experiment we have evaluated the performance of all the methods on a test set of HR images and a test set of LR images. The results of the proposed method (UA-JB), Joint Bayesian (Chen et al, 2012), ITML (Davis et al, 2007) and LMLML (Bohné et al, 2014) are presented in Table 4. UA-JB performs well in all configurations and it worths noticing that, thanks to the use of the uncertainty, it is more robust than other methods.…”

Section: Resolution Changementioning

confidence: 99%

See 1 more Smart Citation

Similarity Function Learning with Data Uncertainty

Bohné

Colin

Gentric

et al. 2016

Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods

Self Cite

View full text Add to dashboard Cite

Similarity functions are at the core of many pattern recognition applications. Standard approaches use feature vectors extracted from a pair of images to compute their degree of similarity. Often feature vectors are noisy and a direct application of standard similarly learning methods may result in unsatisfactory performance. However, information on statistical properties of the feature extraction process may be available, such as the covariance matrix of the observation noise. In this paper, we present a method which exploits this information to improve the process of learning a similarity function. Our approach is composed of an unsupervised dimensionality reduction stage and the similarity function itself. Uncertainty is taken into account throughout the whole processing pipeline during both training and testing. Our method is based on probabilistic models of the data and we propose EM algorithms to estimate their parameters. In experiments we show that the use of uncertainty significantly outperform other standard similarity function learning methods on challenging tasks.

show abstract

Section: Mnistmentioning

confidence: 99%

Section: Resolution Changementioning

confidence: 99%

Similarity Function Learning with Data Uncertainty

Bohné

Colin

Gentric

et al. 2016

Proceedings of the 5th International Conference on Pattern Recognition Applications and Methods

Self Cite

View full text Add to dashboard Cite

show abstract

“…Since discriminative power of input features might vary between different neighbors, learning a global metric may be suboptimal. This has motivated the development of local metric learning approaches [16,1,13,42,2], which increase the discriminative power of global Mahalanobis metric learning by learning a number of local metrics.…”

Section: Local Metric Learningmentioning

confidence: 99%

A Supervised Low-Rank Method for Learning Invariant Subspaces

Siyahjani

Almohsen

Sabri

et al. 2015

2015 IEEE International Conference on Computer Vision (ICCV)

View full text Add to dashboard Cite

Sparse representation and low-rank matrix decomposition approaches have been successfully applied to several computer vision problems. They build a generative representation of the data, which often requires complex training as well as testing to be robust against data variations induced by nuisance factors. We introduce the invariant components, a discriminative representation invariant to nuisance factors, because it spans subspaces orthogonal to the space where nuisance factors are defined. This allows developing a framework based on geometry that ensures a uniform inter-class separation, and a very efficient and robust classification based on simple nearest neighbor. In addition, we show how the approach is equivalent to a local metric learning, where the local metrics (one for each class) are learned jointly, rather than independently, thus avoiding the risk of overfitting without the need for additional regularization. We evaluated the approach for face recognition with highly corrupted training and testing data, obtaining very promising results.

show abstract

“…M (x i , x j ). The definitions of M (x i , x j ), such as k w k (x i , x j )M k in [20] where w k is defined as P (k|x i ) + P (k|x j ) to guarantee the symmetry and P (k|x i ) or P (k|x j ) is based on the posterior probability that the point x belongs to the kth Gaussian cluster in a Gaussian mixture (GMM), are nonetheless not very intuitive.…”

Section: Introductionmentioning

confidence: 99%

Learning Local Metrics and Influential Regions for Classification

Dong

Wang

Yang

et al. 2020

IEEE Trans. Pattern Anal. Mach. Intell.

View full text Add to dashboard Cite

The performance of distance-based classifiers heavily depends on the underlying distance metric, so it is valuable to learn a suitable metric from the data. To address the problem of multimodality, it is desirable to learn local metrics. In this short paper, we define a new intuitive distance with local metrics and influential regions, and subsequently propose a novel local metric learning method for distancebased classification. Our key intuition is to partition the metric space into influential regions and a background region, and then regulate the effectiveness of each local metric to be within the related influential regions. We learn local metrics and influential regions to reduce the empirical hinge loss, and regularize the parameters on the basis of a resultant learning bound. Encouraging experimental results are obtained from various public and popular data sets.

show abstract

Large Margin Local Metric Learning

Cited by 30 publications

References 15 publications

Similarity Function Learning with Data Uncertainty

Similarity Function Learning with Data Uncertainty

A Supervised Low-Rank Method for Learning Invariant Subspaces

Learning Local Metrics and Influential Regions for Classification

Contact Info

Product

Resources

About