We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: Triplet Probabilistic Embedding for Face Verification and Clustering

Abstract: Despite significant progress made over the past twenty five years, unconstrained face verification remains a challenging problem. This paper proposes an approach that couples a deep CNN-based approach with a low-dimensional discriminative embedding learned using triplet probability constraints to solve the unconstrained face verification problem. Aside from yielding performance improvements, this embedding provides significant advantages in terms of memory and for post-processing operations like subject specific clustering. Experiments on the challenging IJB-A dataset show that the proposed algorithm performs comparably or better than the state of the art methods in verification and identification metrics, while requiring much less training data and training time. The superior performance of the proposed method on the CFP dataset shows that the representation learned by our deep CNN is robust to extreme pose variation. Furthermore, we demonstrate the robustness of the deep features to challenges including age, pose, blur and clutter by performing simple clustering experiments on both IJB-A and LFW datasets.
Comments: Oral Paper in BTAS 2016; NVIDIA Best paper Award (this http URL)
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Machine Learning (stat.ML)
DOI: 10.1109/BTAS.2016.7791205
Cite as: arXiv:1604.05417 [cs.CV]
  (or arXiv:1604.05417v3 [cs.CV] for this version)

Submission history

From: Swami Sankaranarayanan [view email]
[v1] Tue, 19 Apr 2016 03:29:56 GMT (4047kb,D)
[v2] Sun, 8 May 2016 16:04:02 GMT (4044kb,D)
[v3] Wed, 18 Jan 2017 03:10:44 GMT (4058kb,D)

Link back to: arXiv, form interface, contact.