Filters








3,980 Hits in 5.8 sec

Deep Metric Learning by Online Soft Mining and Class-Aware Attention [article]

Xinshao Wang, Yang Hua, Elyor Kodirov, Guosheng Hu, Neil M. Robertson
2019 arXiv   pre-print
To address this, we introduce Class-Aware Attention (CAA) that assigns little attention to abnormal data samples.  ...  Deep metric learning aims to learn a deep embedding that can capture the semantic similarity of data points.  ...  Weighted contrastive loss combines Online Soft Mining (OSM) and Class-Aware Attention (CAA) to assign a proper weight for each image pair.Online Soft MiningOnline Soft Mining contains Online Soft Positives  ... 
arXiv:1811.01459v3 fatcat:ntm2v4vjcfdfhmlhhjjkle7y7u

Deep Metric Learning by Online Soft Mining and Class-Aware Attention

Xinshao Wang, Yang Hua, Elyor Kodirov, Guosheng Hu, Neil M. Robertson
2019 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
To address this, we introduce Class-Aware Attention (CAA) that assigns little attention to abnormal data samples.  ...  Deep metric learning aims to learn a deep embedding that can capture the semantic similarity of data points.  ...  Figure 1 : 1 Figure 1: The illustration of Online Soft Mining (OSM) and Class-Aware Attention (CAA) for pair mining. For simplicity, we only consider the scenario with class m and class n.  ... 
doi:10.1609/aaai.v33i01.33015361 fatcat:wklvfundbjd25lb2pthegegzam

Hard-Aware Point-to-Set Deep Metric for Person Re-identification [chapter]

Rui Yu, Zhiyong Dou, Song Bai, Zhaoxiang Zhang, Yongchao Xu, Xiang Bai
2018 Lecture Notes in Computer Science  
To solve this problem, we propose a Hard-Aware Point-to-Set (HAP2S) loss with a soft hard-mining scheme.  ...  Deep metric learning provides a satisfactory solution to person re-ID by training a deep network under supervision of metric loss, e.g., triplet loss.  ...  This work was supported by National Key R&D Program of China No. 2018YFB1004600, NSFC 61703171, and NSFC 61573160, to Dr.  ... 
doi:10.1007/978-3-030-01270-0_12 fatcat:xdz6rgwo6fgfjp6zuqdfqqau5m

Hard-Aware Point-to-Set Deep Metric for Person Re-identification [article]

Rui Yu, Zhiyong Dou, Song Bai, Zhaoxiang Zhang, Yongchao Xu, Xiang Bai
2018 arXiv   pre-print
To solve this problem, we propose a Hard-Aware Point-to-Set (HAP2S) loss with a soft hard-mining scheme.  ...  Deep metric learning provides a satisfactory solution to person re-ID by training a deep network under supervision of metric loss, e.g., triplet loss.  ...  In [18] , a well-designed CNN can learn discriminative features by soft pixel attention and hard regional attention. The loss function in [3] and [18] is softmax loss.  ... 
arXiv:1807.11206v1 fatcat:jahihklmqrhpbiva2w7rhmgd5a

Semantic Granularity Metric Learning for Visual Search [article]

Dipu Manandhar, Muhammet Bastan, Kim-Hui Yap
2019 arXiv   pre-print
Deep metric learning applied to various applications has shown promising results in identification, retrieval and recognition.  ...  then integrate this information into deep metric learning.  ...  , and then integrate this information into deep metric learning.  ... 
arXiv:1911.06047v1 fatcat:njzhrbibtzhgthf3bxfbts67ty

Deep Learning for Person Re-identification: A Survey and Outlook [article]

Mang Ye, Jianbing Shen, Gaojie Lin, Tao Xiang, Ling Shao, Steven C. H. Hoi
2021 arXiv   pre-print
We first conduct a comprehensive overview with in-depth analysis for closed-world person Re-ID from three different perspectives, including deep feature representation learning, deep metric learning and  ...  The widely studied closed-world setting is usually applied under various research-oriented assumptions, and has achieved inspiring success using deep learning techniques on a number of datasets.  ...  A hard-aware point-to-set deep metric learning [155] is designed to mine the informative hard triplets based on the point to set similarity.  ... 
arXiv:2001.04193v2 fatcat:4d3thmsr3va2tnu72nawlu2wxy

Comparing Class-Aware and Pairwise Loss Functions for Deep Metric Learning in Wildlife Re-Identification

Nkosikhona Dlamini, Terence L. van Zyl
2021 Sensors  
Similarity learning using deep convolutional neural networks has been applied extensively in solving computer vision problems.  ...  The advances in similarity learning are essential for smaller datasets or datasets in which few class labels exist per class such as wildlife re-identification.  ...  We also extend our appreciation to Sara Blackburn for allowing us to collect and pre-process the Lion dataset from the living with lions website.  ... 
doi:10.3390/s21186109 pmid:34577319 fatcat:y7d6k7nsb5d7hjpe76rp5mzddy

Deep Metric Learning with Angular Loss [article]

Jian Wang, Feng Zhou, Shilei Wen, Xiao Liu, Yuanqing Lin
2017 arXiv   pre-print
While deep metric learning has yielded impressive performance gains by extracting high level abstractions from image data, a proper objective loss function becomes the central issue to boost the performance  ...  The modern image search system requires semantic understanding of image, and a key yet under-addressed problem is to learn a good metric for measuring the similarity between images.  ...  [14] introduced a position-dependent deep metric unit, which can be used to select hard samples to guide the deep embedding learning in an online and robust manner. More recently, Yuan et al.  ... 
arXiv:1708.01682v1 fatcat:2m553bykirhefpdshm4qkpzhri

Deep Metric Learning to Rank

Fatih Cakir, Kun He, Xide Xia, Brian Kulis, Stan Sclaroff
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
We propose a novel deep metric learning method by revisiting the learning to rank approach.  ...  FastAP has a low complexity compared to existing methods, and is tailored for stochastic gradient descent.  ...  Acknowledgements This work is conducted at Boston University, supported in part by a BU IGNITION award, and equipment donated by NVIDIA.  ... 
doi:10.1109/cvpr.2019.00196 dblp:conf/cvpr/Cakir0XKS19 fatcat:dp7zjz36ovhtlgqhe4zgjchvhm

Sharp Attention Network via Adaptive Sampling for Person Re-identification [article]

Chen Shen, Guo-Jun Qi, Rongxin Jiang, Zhongming Jin, Hongwei Yong, Yaowu Chen, Xian-Sheng Hua
2018 arXiv   pre-print
, Market-1501, and DukeMTMC-reID.  ...  Due to the introduction of sampling-based attention models, the proposed approach can adaptively generate sharper attention-aware feature masks.  ...  ACKNOWLEDGMENT This work was supported in part by the Fundamental Research Funds for the Central Universities.  ... 
arXiv:1805.02336v2 fatcat:uo7ggigxyffrbf6kovda24eweu

Distractor-Aware Deep Regression for Visual Tracking

Ming Du, Yan Ding, Xiuyun Meng, Hua-Liang Wei, Yifan Zhao
2019 Sensors  
In this paper, we propose a novel effective distractor-aware loss function to balance this issue by highlighting the significant domain and by severely penalizing the pure background.  ...  The tracker algorithms directly learn mapping from dense samples around the target object to Gaussian-like soft labels.  ...  The proposed novel distractor-aware loss can alleviate the data-imbalance issue in learning deep-regression networks.  ... 
doi:10.3390/s19020387 fatcat:ocssm6sfofblbgvpoz6okwy6hm

A Metric Learning Reality Check [article]

Kevin Musgrave, Serge Belongie, Ser-Nam Lim
2020 arXiv   pre-print
We find flaws in the experimental methodology of numerous metric learning papers, and show that the actual improvements over time have been marginal at best.  ...  Deep metric learning papers from the past four years have consistently claimed great advances in accuracy, often more than doubling the performance of decade-old methods.  ...  It is no surprise, then, that deep networks have had a similar effect on metric learning.  ... 
arXiv:2003.08505v3 fatcat:soruhibdkjbyrl3wrguua6f7jq

Target-Aware Deep Tracking

Xin Li, Chao Ma, Baoyuan Wu, Zhenyu He, Ming-Hsuan Yang
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
In this paper, we propose a novel scheme to learn target-aware features, which can better recognize the targets undergoing significant appearance variations than pre-trained deep features.  ...  We identify the importance of each convolutional filter according to the back-propagated gradients and select the target-aware features based on activations for representing the targets.  ...  Nam and Han [29] propose a multi-domain deep classifier combined with the hard negative mining, bounding box regression, and online sample collection modules for visual tracking.  ... 
doi:10.1109/cvpr.2019.00146 dblp:conf/cvpr/Li0WH019 fatcat:hjod2ekhzbfjdonfhfxi6iekha

2018 Index IEEE Transactions on Knowledge and Data Engineering Vol. 30

2019 IEEE Transactions on Knowledge and Data Engineering  
Wang, X., Differentially Private Distributed Online Learning. Li, C., þ, TKDE Aug. 2018 1440-1453 FEDERAL: A Framework for Distance-Aware Privacy-Preserving Record Linkage.  ...  Canuto, S., þ, TKDE Dec. 2018 2242-2256 Deep Air Learning: Interpolation, Prediction, and Feature Analysis of Fine-Grained Air Quality.  ... 
doi:10.1109/tkde.2018.2882359 fatcat:asiids266jagrkx5eac6higrlq

Attention-Aware Polarity Sensitive Embedding for Affective Image Retrieval

Xingxu Yao, Dongyu She, Sicheng Zhao, Jie Liang, Yu-Kun Lai, Jufeng Yang
2019 2019 IEEE/CVF International Conference on Computer Vision (ICCV)  
To address the problem, this paper introduces an Attention-aware Polarity Sensitive Embedding (APSE) network to learn affective representations in an end-to-end manner.  ...  Guided by attention module, we weight the sample pairs adaptively which further improves the performance of feature embedding.  ...  Figure 3 . 3 Overview of attention map generation. The class-aware activation and corresponding confidence score are derived in the attention head.  ... 
doi:10.1109/iccv.2019.00123 dblp:conf/iccv/YaoSZLLY19 fatcat:h6fap5mji5e2rm5rddhmucr5qa
« Previous Showing results 1 — 15 out of 3,980 results