Marc Teva Law

Quadruplet-Wise Image Similarity Learning

This paper introduces a novel similarity learning framework. Working with inequality constraints involving quadruplets of images, our approach aims at efficiently modeling similarity from rich or complex semantic label relationships. From these quadruplet-wise constraints, we propose a similarity learning framework relying on a convex optimization scheme. We then study how our metric learning scheme can exploit specific class relationships, such as class ranking (relative attributes), and class taxonomy. We show that classification using the learned metrics gets improved performance over state-of-the-art methods on several datasets. We also evaluate our approach in a new application to learn similarities between web page screenshots in a fully unsupervised way.

Fantope Regularization in Metric Learning

This paper introduces a regularization method to explicitly control the rank of a learned symmetric positive semidefinite distance matrix in distance metric learning. To this end, we propose to incorporate in the objective function a linear regularization term that minimizes the k smallest eigenvalues of the distance matrix. It is equivalent to minimizing the trace of the product of the distance matrix with a matrix in the convex hull of rank-k projection matrices, called a Fantope. Based on this new regularization method, we derive an optimization scheme to efficiently learn the distance matrix. We demonstrate the effectiveness of the method on synthetic and challenging real datasets of face verification and image classification with relative attributes, on which our method outperforms state-of-the-art metric learning algorithms.

Structural and visual similarity learning for Web page archiving

We present in this paper a Web page archiving approach combining image and structural techniques. Our main goal is to learn a similarity between Web pages in order to detect whether successive versions of pages are similar or not. Our system is based on a visual similarity measure designed for Web pages. Combined with a structural analysis of Web page source codes, a supervised feature selection method adapted to Web archiving is proposed. Experiments on real Web archives are reported including scalability issues.

Learning a Distance Metric from Relative Comparisons between Quadruplets of Images

This paper is concerned with the problem of learning a distance metric by considering meaningful and discriminative distance constraints in some contexts where rich information between data is provided. Classic metric learning approaches focus on constraints that involve pairs or triplets of images. We propose a general Mahalanobis-like distance metric learning framework that exploits distance constraints over up to four different images. We show how the integration of such constraints can lead to unsupervised or semi-supervised learning tasks in some applications. We also show the benefit on recognition performance of this type of constraints, in rich contexts such as relative attributes, class taxonomies and temporal webpage analysis.

Structural and visual comparisons for web page archiving

In this paper, we propose a Web page archiving system that combines state-of-the-art comparison methods based on the source codes of Web pages, with computer vision techniques. To detect whether successive versions of a Web page are similar or not, our system is based on: (1) a combination of structural and visual comparison methods embedded in a statistical discriminative model, (2) a visual similarity measure designed for Web pages that improves change detection, (3) a supervised feature selection method adapted to Web archiving. We train a Support Vector Machine model with vectors of similarity scores between successive versions of pages. The trained model then determines whether two versions, defined by their vector of similarity scores, are similar or not. Experiments on real archives validate our approach.

Closed-Form Training of Mahalanobis Distance for Supervised Clustering

Clustering is the task of grouping a set of objects so that objects in the same cluster are more similar to each other than to those in other clusters. The crucial step in most clustering algorithms is to find an appropriate similarity metric, which is both challenging and problem-dependent. Supervised clustering approaches, which can exploit labeled clustered training data that share a common metric with the test set, have thus been proposed. Unfortunately, current metric learning approaches for supervised clustering do not scale to large or even medium-sized datasets. In this paper, we propose a new structured Mahalanobis Distance Metric Learning method for supervised clustering. We formulate our problem as an instance of large margin structured prediction and prove that it can be solved very efficiently in closed-form. The complexity of our method is (in most cases) linear in the size of the training dataset. We further reveal a striking similarity between our approach and multivariate linear regression. Experiments on both synthetic and real datasets confirm several orders of magnitude speedup while still achieving state-of-the-art performance.

Hybrid pooling fusion in the bow pipeline

In the context of object and scene recognition, state-of-the-art performances are obtained with Bag of Words (BoW) models of mid-level representations computed from dense sampled local descriptors (e.g. SIFT). Several methods to combine low-level features and to set mid-level parameters have been evaluated recently for image classification. n nIn this paper, we further investigate the impact of the main parameters in the BoW pipeline. We show that an adequate combination of several low (sampling rate, multiscale) and mid level (codebook size, normalization) parameters is decisive to reach good performances. Based on this analysis, we propose a merging scheme exploiting the specificities of edge-based descriptors. Low and high-contrast regions are pooled separately and combined to provide a powerful representation of images. Sucessful experiments are provided on the Caltech-101 and Scene-15 datasets.

Bag-of-Words Image Representation: Key Ideas and Further Insight

In the context of object and scene recognition, state-of-the-art performances are obtained with visual Bag-of-Words (BoW) models of mid-level representations computed from dense sampled local descriptors (e.g., Scale-Invariant Feature Transform (SIFT)). Several methods to combine low-level features and to set mid-level parameters have been evaluated recently for image classification. In this chapter, we study in detail the different components of the BoW model in the context of image classification. Particularly, we focus on the coding and pooling steps and investigate the impact of the main parameters of the BoW pipeline. We show that an adequate combination of several low (sampling rate, multiscale) and mid-level (codebook size, normalization) parameters is decisive to reach good performances. Based on this analysis, we propose a merging scheme that exploits the specificities of edge-based descriptors. Low and high contrast regions are pooled separately and combined to provide a powerful representation of images. We study the impact on classification performance of the contrast threshold that determines whether a SIFT descriptor corresponds to a low contrast region or a high contrast region. Successful experiments are provided on the Caltech-101 and Scene-15 datasets.