Learning a Distance Metric from Relative Comparisons between Quadruplets of Images

Marc T. Law 1, * Nicolas Thome 1 Matthieu Cord 1
* Corresponding author
1 MLIA - Machine Learning and Information Access
LIP6 - Laboratoire d'Informatique de Paris 6
Abstract : This paper is concerned with the problem of learning a distance metric by considering meaningful and dis-criminative 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.
Document type :
Journal articles
Complete list of metadatas

Cited literature [62 references]  Display  Hide  Download

https://hal.sorbonne-universite.fr/hal-01346190
Contributor : Gestionnaire Hal-Upmc <>
Submitted on : Monday, July 18, 2016 - 2:35:43 PM
Last modification on : Friday, May 24, 2019 - 5:31:05 PM

File

Law_2016_Learning_a_Distance.p...
Files produced by the author(s)

Identifiers

Citation

Marc T. Law, Nicolas Thome, Matthieu Cord. Learning a Distance Metric from Relative Comparisons between Quadruplets of Images. International Journal of Computer Vision, Springer Verlag, 2016, pp.1-30. ⟨10.1007/s11263-016-0923-4⟩. ⟨hal-01346190⟩

Share

Metrics

Record views

274

Files downloads

286