Filters








10,520 Hits in 4.9 sec

Multi-level Semantic Feature Augmentation for One-shot Learning [article]

Zitian Chen, Yanwei Fu, Yinda Zhang, Yu-Gang Jiang, Xiangyang Xue, and Leonid Sigal
2019 arXiv   pre-print
The encoder part of the TriNet learns to map multi-layer visual features of deep CNNs, that is, multi-level concepts, to a semantic vector.  ...  In this paper, we propose a novel approach to one-shot learning that builds on this idea.  ...  Dual TriNet learns the mapping between the Multi-level Image Feature (M-IF) encoding and the Semantic space.  ... 
arXiv:1804.05298v4 fatcat:ce4qzqpstfhdhfhqkk2nqtwbhe

Semantic Embedding Space for Zero-Shot Action Recognition [article]

Xun Xu, Timothy Hospedales, Shaogang Gong
2015 arXiv   pre-print
This is more challenging because the mapping between the semantic space and space-time features of videos containing complex actions is more complex and harder to learn.  ...  The number of categories for action recognition is growing rapidly. It is thus becoming increasingly hard to collect sufficient training data to learn conventional models for each category.  ...  By augmenting target training data X trg tr with auxiliary data X aux tr we can achieve the state-of-the-art performance for Zero-shot Learning and competitive performance on Multi-shot Learning data X  ... 
arXiv:1502.01540v1 fatcat:lsmmzc57lfgwthagdj2ba2ig3i

Semantic embedding space for zero-shot action recognition

Xun Xu, Timothy Hospedales, Shaogang Gong
2015 2015 IEEE International Conference on Image Processing (ICIP)  
This is more challenging because the mapping between the semantic space and space-time features of videos containing complex actions is more complex and harder to learn.  ...  The number of categories for action recognition is growing rapidly. It is thus becoming increasingly hard to collect sufficient training data to learn conventional models for each category.  ...  By augmenting target training data X trg tr with auxiliary data X aux tr we can achieve the state-of-the-art performance for Zero-shot Learning and competitive performance on Multi-shot Learning data X  ... 
doi:10.1109/icip.2015.7350760 dblp:conf/icip/XuHG15 fatcat:xksmmblsq5dbloutbowq7f34iq

Rethinking Class Relations: Absolute-relative Supervised and Unsupervised Few-shot Learning

Hongguang Zhang, Piotr Koniusz, Songlei Jian, Hongdong Li, Philip H. S. Torr
2021 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
For instance, children learn the concept of tiger from a few of actual examples as well as from comparisons of tiger to other animals.  ...  The majority of existing few-shot learning methods describe image relations with binary labels.  ...  In contrast, we focus on how to refine the backbone by learning from class concepts and relations to address the high-level few-shot learning task.  ... 
doi:10.1109/cvpr46437.2021.00931 fatcat:5wzermromrdi5gfwq6uhtway54

A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunities [article]

Yisheng Song, Ting Wang, Subrota K Mondal, Jyoti Prakash Sahoo
2022 arXiv   pre-print
For the sake of avoiding conceptual confusion, we first elaborate and compare a set of similar concepts including few-shot learning, transfer learning, and meta-learning.  ...  Few-shot learning (FSL) has emerged as an effective learning method and shows great potential.  ...  Approach Representation Alignment Fusion Co-Learning Translation  ... 
arXiv:2205.06743v2 fatcat:xmxht2ileja53o2o5b4vrw32ey

Knowledge Extraction in Low-Resource Scenarios: Survey and Perspective [article]

Shumin Deng, Ningyu Zhang, Hui Chen, Feiyu Xiong, Jeff Z. Pan, Huajun Chen
2022 arXiv   pre-print
Many neural approaches on low-resource KE have been widely investigated and achieved impressive performance.  ...  In addition, we describe promising applications and outline some potential directions for future research.  ...  For few-shot RE, [Gao et al., 2019a] have equipped prototypical network with hybrid attention which consists of an instance-level attention and a feature-level attention to select more informative instances  ... 
arXiv:2202.08063v1 fatcat:2q64tx2mzne53gt24adi6ymj7a

Learning Implicit Temporal Alignment for Few-shot Video Classification [article]

Songyang Zhang, Jiale Zhou, Xuming He
2021 arXiv   pre-print
To train our model, we develop a multi-task loss for learning video matching, leading to video features with better generalization.  ...  To address this, we propose a novel matching-based few-shot learning strategy for video sequences in this work.  ...  To perform few-shot learning, we develop a meta-learning strategy that employs a multi-task loss to simultaneously exploit the meta-level and semantic-level supervision.  ... 
arXiv:2105.04823v1 fatcat:tthebmigibf7rirkt52uvna4ee

Multi-Task Zero-Shot Action Recognition with Prioritised Data Augmentation [chapter]

Xun Xu, Timothy M. Hospedales, Shaogang Gong
2016 Lecture Notes in Computer Science  
This is achieved by establishing a mapping connecting low-level features and a semantic description of the label space, referred as visual-semantic mapping, on auxiliary data.  ...  Zero-Shot Learning (ZSL) promises to scale visual recognition by bypassing the conventional model training requirement of annotated examples for every category.  ...  A mapping (e.g. regression [9] or bilinear model [7] ) is learned between low-level visual features and their semantic embeddings.  ... 
doi:10.1007/978-3-319-46475-6_22 fatcat:zmwhktds3vfndj6wh364qpsutu

A Transductive Multi-Head Model for Cross-Domain Few-Shot Learning [article]

Jianan Jiang, Zhenpeng Li, Yuhong Guo, Jieping Ye
2020 arXiv   pre-print
In this paper, we present a new method, Transductive Multi-Head Few-Shot learning (TMHFS), to address the Cross-Domain Few-Shot Learning (CD-FSL) challenge.  ...  For the few-shot learning in the target domain, we first perform fine-tuning on the embedding network with only the semantic global classifier and the support instances, and then use the MCT part to predict  ...  Figure 1 . 1 The proposed transductive multi-head few-shot (TMHFS) learning model. few-shot learning evaluation.  ... 
arXiv:2006.11384v1 fatcat:qine46e72ngltb25wcxipnmhqy

LaSO: Label-Set Operations Networks for Multi-Label Few-Shot Learning

Amit Alfassy, Leonid Karlinsky, Amit Aides, Joseph Shtok, Sivan Harary, Rogerio Feris, Raja Giryes, Alex M. Bronstein
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
context of performing data augmentation for multi-label few-shot learning.  ...  This enables performing augmentation on examples of novel categories, thus, facilitating multi-label few-shot classifier learning.  ...  than on the feature level, as in our approach.  ... 
doi:10.1109/cvpr.2019.00671 dblp:conf/cvpr/AlfassyKASHFGB19 fatcat:achokx7yxnhgtl3yqsf6v5s4au

Deep Zero-Shot Learning for Scene Sketch [article]

Yao Xie and Peng Xu and Zhanyu Ma
2019 arXiv   pre-print
network for scene sketch feature learning.  ...  To overcome these challenges, we propose a deep embedding model for scene sketch zero-shot learning.  ...  Based on our augmented semantic vector, we propose a deep embedding model to solve scene sketch zero-shot learning, in which we adopt visual feature space of scene sketch as the embedding space to alleviate  ... 
arXiv:1905.04510v1 fatcat:wkfghirrtneilh24rnrb6ouzwm

LaSO: Label-Set Operations networks for multi-label few-shot learning [article]

Amit Alfassy, Leonid Karlinsky, Amit Aides, Joseph Shtok, Sivan Harary, Rogerio Feris, Raja Giryes, Alex M. Bronstein
2019 arXiv   pre-print
context of performing data augmentation for multi-label few-shot learning.  ...  This enables performing augmentation on examples of novel categories, thus, facilitating multi-label few-shot classifier learning.  ...  than on the features level, which we aim at.  ... 
arXiv:1902.09811v1 fatcat:lt7rnqruibd3loy7ifcafdlwdy

Part-aware Prototype Network for Few-shot Semantic Segmentation [article]

Yongfei Liu, Xiangyi Zhang, Songyang Zhang, Xuming He
2020 arXiv   pre-print
Few-shot semantic segmentation aims to learn to segment new object classes with only a few annotated examples, which has a wide range of real-world applications.  ...  In this paper, we propose a novel few-shot semantic segmentation framework based on the prototype representation.  ...  In order to learn better visual representation for few-shot segmentation, we introduce another semantic branch [34] for computing a semantic loss defined on the global semantic class space C tr (in contrast  ... 
arXiv:2007.06309v2 fatcat:cqjjufddlbcihds4zxcpxlr7vu

Few-Shot Semantic Segmentation Augmented with Image-Level Weak Annotations [article]

Shuo Lei, Xuchao Zhang, Jianfeng He, Fanglan Chen, Chang-Tien Lu
2021 arXiv   pre-print
However, these few-shot approaches cannot easily be applied to multi-way or weak annotation settings.  ...  Recent progress in fewshot semantic segmentation tackles the issue by only a few pixel-level annotated examples.  ...  CONCLUSION In this paper, a novel weak-annotation-augmented few-shot segmentation model is proposed to learn an augmented prototype based on both pixel-level and image-level annotations.  ... 
arXiv:2007.01496v2 fatcat:aeal67icjfazbcrpos7qybbolm

Learning Better Features for Face Detection with Feature Fusion and Segmentation Supervision [article]

Wanxin Tian, Zixuan Wang, Haifeng Shen, Weihong Deng, Yiping Meng, Binghui Chen, Xiubao Zhang, Yuan Zhao, Xiehe Huang
2019 arXiv   pre-print
Specifically, inspired by FPN and SENet, we apply semantic information from higher-level feature maps as contextual cues to augment low-level feature maps via a spatial and channel-wise attention style  ...  We assume that problems inside are inadequate use of supervision information and imbalance between semantics and details at all level feature maps in CNN even with Feature Pyramid Networks (FPN).  ...  to augment semantics in lower-level feature maps in a spatial and channel-wise attention manner. • We improve the typical deep single shot detectors by making up for anchor mechanism with a semantic segmentation  ... 
arXiv:1811.08557v3 fatcat:h7vo3jaw6vginntlxqzyhlmhsi
« Previous Showing results 1 — 15 out of 10,520 results