Filters








153,406 Hits in 1.8 sec

Variational Continual Proxy-Anchor for Deep Metric Learning

Minyoung Kim, Ricardo Guerrero, Hai Xuan Pham, Vladimir Pavlovic
2022 International Conference on Artificial Intelligence and Statistics  
The recent proxy-anchor method achieved outstanding performance in deep metric learning, which can be acknowledged to its data efficient loss based on hard example mining, as well as far lower sampling  ...  By regarding each batch as a task in continual learning, we adopt the Bayesian variational continual learning approach to derive a novel loss function.  ...  INTRODUCTION Deep metric learning (DML) is the task of learning a metric (similarity or distance measure) between two data points (e.g., images), which can be useful for numerous downstream computer vision  ... 
dblp:conf/aistats/KimGPP22 fatcat:fxdgrudagnec7jznd2suv5rgne

High-Dimensional Bayesian Optimisation with Variational Autoencoders and Deep Metric Learning [article]

Antoine Grosnit, Rasul Tutunov, Alexandre Max Maraval, Ryan-Rhys Griffiths, Alexander I. Cowen-Rivers, Lin Yang, Lin Zhu, Wenlong Lyu, Zhitang Chen, Jun Wang, Jan Peters, Haitham Bou-Ammar
2021 arXiv   pre-print
We introduce a method combining variational autoencoders (VAEs) and deep metric learning to perform Bayesian optimisation (BO) over high-dimensional and structured input spaces.  ...  By adapting ideas from deep metric learning, we use label guidance from the blackbox function to structure the VAE latent space, facilitating the Gaussian process fit and yielding improved BO performance  ...  and deep metric learning.  ... 
arXiv:2106.03609v3 fatcat:hzy5d6iawfhdna2qpbx3w57mx4

Facial Kinship Verification with Large Age Variation Using Deep Linear Metric Learning

Diego Lelis, Department of Mechanical Engineering, University of Brasilia, Brasilia, Brazil, Dibio L. Borges
2019 Journal of Image and Graphics  
This research proposed a solution to the kinship verification problem with a novel non-context-aware approach using a dataset with large age variation by applying our proposed method Deep Linear Metric  ...  Our method leverages multiple deep learning architectures trained with massive facial datasets.  ...  Díbio Leandro Borges, for all the guidance an support throughout this research, to the University of Brasília(UnB) that provided me with an exceptional environment for research and learning.  ... 
doi:10.18178/joig.7.2.50-58 fatcat:wzhhk35ur5bdtmxo5hp4sdab2m

Variational Autoencoders For Semi-Supervised Deep Metric Learning

Nathan Safir, Meekail Zain, Curtis Godwin, Eric Miller, Bella Humphrey, Shannon Quinn
2022 Proceedings of the Python in Science Conferences   unpublished
Deep metric learning (DML) methods generally do not incorporate unlabelled data.  ...  We propose borrowing components of the variational autoencoder (VAE) methodology to extend DML methods to train on semi-supervised datasets.  ...  Related Literature The goal of this research is to investigate how components of the variational autoencoder can help the performance of deep metric learning in semi supervised tasks.  ... 
doi:10.25080/majora-212e5952-022 fatcat:t2n7pjwkw5buxiou4mcywejegi

TVAE: Triplet-Based Variational Autoencoder using Metric Learning [article]

Haque Ishfaq, Assaf Hoogi, Daniel Rubin
2018 arXiv   pre-print
In this project, we propose a novel integrated framework to learn latent embedding in VAE by incorporating deep metric learning.  ...  from metric learning.  ...  approximate inference that leverage both traditional VAE and deep metric learning techniques.  ... 
arXiv:1802.04403v2 fatcat:htxs437c6jcanlwmc35ob4ziwu

Embedding Deep Metric for Person Re-identication A Study Against Large Variations [article]

Hailin Shi, Yang Yang, Xiangyu Zhu, Shengcai Liao, Zhen Lei, Weishi Zheng, Stan Z. Li
2016 arXiv   pre-print
In addition, we improve the learning by a metric weight constraint, so that the learned metric has a better generalization ability.  ...  Experiments show that these two strategies are effective in learning robust deep metrics for person re-identification, and accordingly our deep model significantly outperforms the state-of-the-art methods  ...  Weight Constraint for Metric Learning. A commonly used metric by deep learning methods is the Euclidean distance [6, 30, 27] .  ... 
arXiv:1611.00137v1 fatcat:bks6hcgzhvfhlj77xvo76h7yem

Video-Based Person Re-Identification: Methods, Datasets, and Deep Learning

2020 International Journal of Engineering and Advanced Technology  
Feature representation and metric learning are major issues for person re-identification.  ...  The last decade witnessed the emergence of large-scale datasets and deep learning methods to use these huge data volumes.  ...  Model learning has to be generalized with optimal metrics. Deep learning of features and variations is coming up to address the issue up to a certain extent. V.  ... 
doi:10.35940/ijeat.c6524.029320 fatcat:zbaiu3k7yncc3djk3onwf2jklq

Enhancing Multi-tissue and Multi-scale Cell Nuclei Segmentation with Deep Metric Learning

Tomas Iesmantas, Agne Paulauskaite-Taraseviciene, Kristina Sutiene
2020 Applied Sciences  
4) Conclusion: We conclude that deep metric learning gives an additional boost to the overall learning process and consequently improves the segmentation performance.  ...  Notably, the improvement ranges approximately between 0.13% and 22.31% for different types of images in the terms of Dice coefficients when compared to no metric deep learning.  ...  Discussion The concept of deep metric learning was introduced when deep learning and metric learning were combined [34] .  ... 
doi:10.3390/app10020615 fatcat:pgxqbut5f5bp7ntszh3cvhxipq

Deep adaptive feature embedding with local sample distributions for person re-identification

Lin Wu, Yang Wang, Junbin Gao, Xue Li
2018 Pattern Recognition  
To this end, a novel objective function is proposed to jointly optimize similarity metric learning, local positive mining and robust deep embedding.  ...  Our method is capable of learning a deep similarity metric adaptive to local sample structure by minimizing each sample's local distances while propagating through the relationship between samples to attain  ...  Inspired by these high-capacity models in deep learning, some deep embedding models have been developed for person re-id to learn representations against visual variations [19, 20, 21, 22, 23, 13, 24,  ... 
doi:10.1016/j.patcog.2017.08.029 fatcat:xt65d6molnhbfpriyvjskyjeym

Deep Metric Learning with Density Adaptivity [article]

Yehao Li and Ting Yao and Yingwei Pan and Hongyang Chao and Tao Mei
2019 arXiv   pre-print
With the rise and success of Convolutional Neural Networks (CNN), deep metric learning (DML) involves training a network to learn a nonlinear transformation to the embedding space.  ...  The problem of distance metric learning is mostly considered from the perspective of learning an embedding space, where the distances between pairs of examples are in correspondence with a similarity metric  ...  https://github.com/rksltnl/Deep-Metric-Learning-CVPR16/tree/ master/code/evaluation  ... 
arXiv:1909.03909v1 fatcat:b4dj5qns3jendd2n2n2cx5nquq

Exploring Intensity Invariance in Deep Neural Networks for Brain Image Registration [article]

Hassan Mahmood, Asim Iqbal, Syed Mohammed Shamsul Islam
2020 arXiv   pre-print
In this study, we investigate the effect of variation in intensity distribution among input image pairs for deep learning-based image registration methods.  ...  However, deep learning-based techniques are shown to be computationally efficient for automated brain registration but are sensitive to the intensity variations.  ...  networks as well as introducing structural similarity metric in a loss function for deep learning-based image registration techniques.  ... 
arXiv:2009.10058v1 fatcat:gsbfqnwlf5dhho64azoca2rf3y

Memory based neural networks for end-to-end autonomous driving [article]

Sergio Paniego Blanco, Sakshay Mahna, Utkarsh A. Mishra, JoseMaria Canas
2022 arXiv   pre-print
We describe and compare this architecture with previous approaches using fundamental error metrics (MAE, MSE) and several external metrics based on their performance on simulated test circuits.  ...  and its variations, improving the performance and generalization of networks that are memory-less and approaches that are not based on deep learning.  ...  Memory-based deep learning models Two models are presented and tested as memory-based options.  ... 
arXiv:2205.12124v1 fatcat:uk5vhge2rvgarkqsj7x7zox76u

Constrained Deep Metric Learning for Person Re-identification [article]

Hailin Shi and Xiangyu Zhu and Shengcai Liao and Zhen Lei and Yang Yang and Stan Z. Li
2015 arXiv   pre-print
Firstly, a novel deep architecture is built where the Mahalanobis metric is learned with a weight constraint.  ...  This weight constraint is used to regularize the learning, so that the learned metric has a better generalization ability.  ...  Fig. 2 is an overview of the network for Constrained Deep Metric Learning (CDML).  ... 
arXiv:1511.07545v1 fatcat:4xfcilg36bh67j5n44k3etgq6m

Bayesian Deep Neural Networks for Supervised Learning of Single-View Depth [article]

Javier Rodríguez-Puigvert, Rubén Martínez-Cantín, Javier Civera
2021 arXiv   pre-print
In this paper, we evaluate scalable approaches to uncertainty quantification in single-view supervised depth learning, specifically MC dropout and deep ensembles.  ...  We show that adding dropout in all layers of the encoder brings better results than other variations found in the literature.  ...  Table II shows the depth and uncertainty metrics for the MC dropout variations and deep ensembles on SceneNet RGB-D.  ... 
arXiv:2104.14202v3 fatcat:5nnsa5jbn5c37po722kau3y4wm

Nonlinear Local Metric Learning for Person Re-identification [article]

Siyuan Huang, Jiwen Lu, Jie Zhou, Anil K. Jain
2015 arXiv   pre-print
of samples, we utilize the merits of both local metric learning and deep neural network to learn multiple sets of nonlinear transformations.  ...  A discriminative metric learning method should be capable of exploiting complex nonlinear transformations due to the large variations in feature space.  ...  ], large margin nearest neighbor (LMNN) [44] , and information theoretic metric learning (ITML) [9] , discriminative deep metric learning (DDML) [20] and large margin local metric learning (LMLML)  ... 
arXiv:1511.05169v1 fatcat:bigdznbdoveyzksjajrlvkrkka
« Previous Showing results 1 — 15 out of 153,406 results