Deep Metric Learning via Lifted Structured Feature Embedding

Song, Hyun Oh; Xiang, Yu; Jegelka, Stefanie; Savarese, Silvio

doi:10.1109/cvpr.2016.434

Cited by 1,430 publications

(1,544 citation statements)

References 28 publications

(72 reference statements)

Supporting

Mentioning

1,493

Contrasting

Unclassified

Order By: Relevance

“…Recently, there have been a lot of methods to add the verification information to the CNN for face verification task, such as contrastive loss [3], triplet loss [1], Lifted structured embedding [6]. The CNN trained with verification information can adjust the parameters end-to-end, so that the features generated from these CNN have greater discriminant power than those from normal networks which just use the cross entropy loss for objective.…”

Section: The Proposed Loss Functionmentioning

confidence: 99%

“…Triplet loss [1] proposed online and offline methods for selecting training pairs, and each anchor uses semi-hard sample as its corresponding negative sample. Although Lifted structured embedding [6] does not need to pair the samples in a complex method, if the batchsize is N, a high cost O(N 2 ) is entailed. The research community still does not have reasonable ways to pair samples.…”

Section: The Proposed Loss Functionmentioning

confidence: 99%

See 1 more Smart Citation

Face Verification with Multi-Task and Multi-Scale Features Fusion

Lu¹,

Yang²,

Zhang³

et al. 2017

Preprint

View full text Add to dashboard Cite

Abstract:Face verification for unrestricted faces in the wild is a challenging task. This paper proposes a method based on two deep convolutional neural networks(CNN) for face verification. In this work, we explore to use identification signal to supervise one CNN and the combination of semi-verification and identification to train the other one. In order to estimate semi-verification loss at a low computation cost, a circle, which is composed of all faces, is used for selecting face pairs from pairwise samples. In the process of face normalization, we propose to use different landmarks of faces to solve the problems caused by poses. And the final face representation is formed by the concatenating feature of each deep CNN after PCA reduction. What's more, each feature is a combination of multi-scale representations through making use of auxiliary classifiers. For the final verification, we only adopt the face representation of one region and one resolution of a face jointing Joint Bayesian classifier. Experiments show that our method can extract effective face representation with a small training dataset and our algorithm achieves 99.71% verification accuracy on LFW dataset.

show abstract

Section: The Proposed Loss Functionmentioning

confidence: 99%

Section: The Proposed Loss Functionmentioning

confidence: 99%

Face Verification with Multi-Task and Multi-Scale Features Fusion

Lu¹,

Yang²,

Zhang³

et al. 2017

Preprint

View full text Add to dashboard Cite

show abstract

“…Before each CSS layer, convolutional activations are normalized to have a L 2 norm [43]. To learn the network, we employed the Caltech-101 dataset [13] excluding testing image pairs used in experiments.…”

Section: Experimental Settingsmentioning

confidence: 99%

FCSS: Fully Convolutional Self-Similarity for Dense Semantic Correspondence

Kim

Min

Ham

et al. 2019

IEEE Trans. Pattern Anal. Mach. Intell.

View full text Add to dashboard Cite

We present a descriptor, called fully convolutional selfsimilarity (FCSS), for dense semantic correspondence. To robustly match points among different instances within the same object class, we formulate FCSS using local selfsimilarity (LSS) within a fully convolutional network. In contrast to existing CNN-based descriptors, FCSS is inherently insensitive to intra-class appearance variations because of its LSS-based structure, while maintaining the precise localization ability of deep neural networks. The sampling patterns of local structure and the self-similarity measure are jointly learned within the proposed network in an end-to-end and multi-scale manner. As training data for semantic correspondence is rather limited, we propose to leverage object candidate priors provided in existing image datasets and also correspondence consistency between object pairs to enable weakly-supervised learning. Experiments demonstrate that FCSS outperforms conventional handcrafted descriptors and CNN-based descriptors on various benchmarks.

show abstract

“…Comparatively, stochastic triple embedding [24] learns the representation in a triplet manner, and joint Bayesian model (JBM) starts from a different assumption to obtain success in face verification task [25]. Magnet loss (MNL) [13] and lifted structured feature embedding [26] are two latest well-designed methods for enhanced local discrimination modeling. In ASV scenarios, NDA has proved to be a strong substitute for LDA [18].…”

Section: Distance Metric Learningmentioning

confidence: 99%

Centroid-aware local discriminative metric learning in speaker verification

Sheng

Dong

et al. 2017

Pattern Recognition

View full text Add to dashboard Cite

We propose a new mechanism to pave the way for efficient learning against class-imbalance and improve representation of identity vector (i-vector) in automatic speaker verification (ASV). The insight is to effectively exploit the inherent structure within ASV corpus -centroid priori. In particular: 1) to ensure learning efficiency against class-imbalance, the centroid-aware balanced boosting sampling is proposed to collect balanced mini-batch; 2) to strengthen local discriminative modeling on the mini-batches, neighborhood component analysis (NCA) and magnet loss (MNL) are adopted in ASV-specific modifications. The integration creates adaptive NCA (AdaNCA) and linear MNL (LMNL). Numerical results show that LMNL is a competitive candidate for lowdimensional projection on i-vector (EER=3.84% on SRE2008, EER=1.81% on SRE2010), enjoying competitive edge over linear discriminant analysis (LDA). AdaNCA (EER=4.03% on SRE2008, EER=2.05% on SRE2010) also performs well. Furthermore, to facilitate the future study on boosting sampling, connections between boosting sampling, hinge loss and data augmentation have been established, which help understand the behavior of boosting sampling further.

show abstract

Deep Metric Learning via Lifted Structured Feature Embedding

Cited by 1,430 publications

References 28 publications

Face Verification with Multi-Task and Multi-Scale Features Fusion

Face Verification with Multi-Task and Multi-Scale Features Fusion

FCSS: Fully Convolutional Self-Similarity for Dense Semantic Correspondence

Centroid-aware local discriminative metric learning in speaker verification

Contact Info

Product

Resources

About