32,893 Hits in 6.4 sec

Semi-supervised learning of object categories from paired local features

Wen Wu, Jie Yang
2008 Proceedings of the 2008 international conference on Content-based image and video retrieval - CIVR '08  
This paper presents a semi-supervised learning (SSL) approach to find similarities of images using statistics of local matches.  ...  Our experiments confirm that our SSL based approach not only boost classification performance but also improve robustness of the learned category model using only simple local keypoint features.  ...  Figure 1 : 1 Two kinds of image matching: object based matching (a) and scene based matching (b). Figure 2 : 2 Semi-supervised learning (SSL) of the airplane and motorbike categories.  ... 
doi:10.1145/1386352.1386386 dblp:conf/civr/WuY08 fatcat:4bse67eiyfajndeqyhcr665ssi

Reconstruction-Aware Prior Distillation for Semi-supervised Point Cloud Completion [article]

Zhaoxin Fan, Yulin He, Zhicheng Wang, Kejian Wu, Hongyan Liu, Jun He
2022 arXiv   pre-print
a small number of paired training samples.  ...  In training stage 1, the so-called deep semantic prior is learned from both unpaired complete and unpaired incomplete point clouds using a reconstruction-aware pretraining process.  ...  Due to its disability in extracting local feature, PCN fails to recover the geometric details of the object.  ... 
arXiv:2204.09186v2 fatcat:g6cuoay72fcwhnwjqupnj7rwu4

Local-driven semi-supervised learning with multi-label

Teng Li, Shuicheng Yan, Tao Mei, In-So Kweon
2009 2009 IEEE International Conference on Multimedia and Expo  
The experiments on multi-label image annotation demonstrate the encouraging results from our proposed framework of semi-supervised learning.  ...  In this paper, we present a local-driven semi-supervised learning framework to propagate the labels of the training data (with multi-label) to the unlabeled data.  ...  Building Graph Based on Local Matches We extract two types of local features: Scale-Invariant Feature Transform (SIFT) local features [10] for object categories such as "building", "animal", and segmented  ... 
doi:10.1109/icme.2009.5202790 dblp:conf/icmcs/LiYMK09 fatcat:sdp3q37elra3tjk7dbkzeolguq

GuidedMix-Net: Learning to Improve Pseudo Masks Using Labeled Images as Reference [article]

Peng Tu, Yawen Huang, Rongrong Ji, Feng Zheng, Ling Shao
2021 arXiv   pre-print
Semi-supervised learning is a challenging problem which aims to construct a model by learning from a limited number of labeled examples.  ...  Specifically, we first introduce a feature alignment objective between labeled and unlabeled data to capture potentially similar image pairs and then generate mixed inputs from them.  ...  Then, we learn the uniform feature vectors from the mixed data to inherit different contexts from the image pairs.  ... 
arXiv:2106.15064v2 fatcat:6z3i3lx7f5ce3lcqhbafxbgwua

Guided Point Contrastive Learning for Semi-supervised Point Cloud Semantic Segmentation [article]

Li Jiang, Shaoshuai Shi, Zhuotao Tian, Xin Lai, Shu Liu, Chi-Wing Fu, Jiaya Jia
2021 arXiv   pre-print
Also, we design the confidence guidance to ensure high-quality feature learning.  ...  Inspired by the recent contrastive loss in self-supervised tasks, we propose the guided point contrastive loss to enhance the feature representation and model generalization ability in semi-supervised  ...  Acknowledgments The project is supported in part by the Research Grants Council of the Hong Kong Special Administrative Region (Project no. CUHK 14206320).  ... 
arXiv:2110.08188v1 fatcat:i25xkcemj5hopn6pgicf4infue

Category Modeling from Just a Single Labeling: Use Depth Information to Guide the Learning of 2D Models

Quanshi Zhang, Xuan Song, Xiaowei Shao, Ryosuke Shibasaki, Huijing Zhao
2013 2013 IEEE Conference on Computer Vision and Pattern Recognition  
We design a graphical model that uses object edges to represent object structures, and this paper aims to incrementally learn this category model from one labeled object and a number of casually captured  ...  An object model base that covers a large number of object categories is of great value for many computer vision tasks.  ...  Second, if we idealize the spirit of semi-supervised learning, can we learn a category model from the minimum labeling (only one labeled object) and casually captured image sample pools?  ... 
doi:10.1109/cvpr.2013.32 dblp:conf/cvpr/ZhangSSSZ13 fatcat:kjcivhxbgncnjdiztbrig6b32q

Leveraging Unlabeled Data for Sketch-based Understanding [article]

Javier Morales, Nils Murrugarra-Llerena, Jose M. Saavedra
2022 arXiv   pre-print
To this end, we evaluate variations of VAE and semi-supervised VAE, and present an extension of BYOL to deal with sketches.  ...  Our results show the superiority of sketch-BYOL, which outperforms other self-supervised approaches increasing the retrieval performance for known and unknown categories.  ...  Sketch-based localization The idea for a model is to localize all instances of an object in a regular image (scene). A sketch represents the target object.  ... 
arXiv:2204.12522v1 fatcat:uj4osg4wlja3bczob7bplb4k2u

Learning to Localize Sound Source in Visual Scenes [article]

Arda Senocak, Tae-Hyun Oh, Junsik Kim, Ming-Hsuan Yang, In So Kweon
2018 arXiv   pre-print
Moreover, although our network is formulated within the unsupervised learning framework, it can be extended to a unified architecture with a simple modification for the supervised and semi-supervised learning  ...  We pose the question: Can the machine learn the correspondence between visual scene and the sound, and localize the sound source only by observing sound and visual scene pairs like human?  ...  We are grateful to the annotators for localizing the sound sources on our dataset.  ... 
arXiv:1803.03849v1 fatcat:vubtdsrybnayxizc4wxwvd2kdi

Instant Response Few-shot Object Detection with Meta Strategy and Explicit Localization Inference [article]

Junying Huang, Fan Chen, Sibo Huang, Dongyu Zhang
2022 arXiv   pre-print
Aiming at recognizing and localizing the object of novel categories by a few reference samples, few-shot object detection (FSOD) is a quite challenging task.  ...  Then, we introduce two explicit inferences into the localization module to alleviate its over-fitting to the base categories, including explicit localization score and semi-explicit box regression.  ...  ., dynamic classifier and semi-supervised RPN (SS-RPN), to enable the object detection of novel categories with instant response.  ... 
arXiv:2110.13377v2 fatcat:ntauaytv7rbubpoqdo7esmzxru

Discriminative Mixture-of-Templates for Viewpoint Classification [chapter]

Chunhui Gu, Xiaofeng Ren
2010 Lecture Notes in Computer Science  
A large number of components are learned in the mixture and they are associated with canonical viewpoints of the object through different levels of supervision, being fully supervised, semi-supervised,  ...  In addition, the mixture-of-templates approach to object viewpoint/pose has a natural extension to the continuous case by discriminatively learning a linear appearance model locally at each discrete view  ...  With the maturation of local feature detection (as in SIFT and its variants), latest progresses on pose estimation have mostly been local-feature based (e.g.  ... 
doi:10.1007/978-3-642-15555-0_30 fatcat:mfvypwb7iff2beaxdynayg3ily

Foreground Focus: Finding Meaningful Features in Unlabeled Images

Y.J. Lee, K. Grauman
2008 Procedings of the British Machine Vision Conference 2008  
We show that this mutual reinforcement of object-level and feature-level similarity improves unsupervised image clustering, and apply the technique to automatically discover categories and foreground regions  ...  We present a method to automatically discover meaningful features in unlabeled image collections. Each image is decomposed into semi-local features that describe neighborhood appearance and geometry.  ...  Related Work In this section we review relevant work in supervised image feature selection, weakly supervised and unsupervised category learning, and semi-local descriptors.  ... 
doi:10.5244/c.22.52 dblp:conf/bmvc/LeeG08 fatcat:cpnpmyn3p5euxlyqftqvjh7v54

Semi-Supervised Locality Discriminant Projection

Yu Mao, Yanquan Zhou, Hao Yu, Li Wei, Xiaojie Wang
2012 Procedia Engineering  
In this paper, we consider the problem of semi-supervised dimensionality reduction.  ...  We focus on the local geometric structure of data and propose a novel method, called Semi-supervised Locality Discriminant Projections (SSLDP). It uses both labeled and unlabeled samples.  ...  Acknowledgements This paper was supported by Mechanism socialist method and higher intelligence theory of the national natural science fund projects (No. 60873001).  ... 
doi:10.1016/j.proeng.2012.01.134 fatcat:yqbtbvuvzrghhojcwgluet332m

Generalized Product Quantization Network for Semi-supervised Image Retrieval [article]

Young Kyun Jang, Nam Ik Cho
2020 arXiv   pre-print
To resolve this issue, we propose the first quantization-based semi-supervised image retrieval scheme: Generalized Product Quantization (GPQ) network.  ...  Image retrieval methods that employ hashing or vector quantization have achieved great success by taking advantage of deep learning.  ...  In this section, we will describe each component and how GPQ is learned in a semi-supervised way. Semi-Supervised Learning The feature extractor F generates D-dimensional feature vectorx ∈ R D .  ... 
arXiv:2002.11281v3 fatcat:v27lks4kvffbrkviy7ubveuebi

Semi-Supervised Co-Analysis of 3D Shape Styles from Projected Lines

Fenggen Yu, Yan Zhang*, Kai Xu, Ali Mahdavi-Amiri, Hao Zhang
2018 ACM Transactions on Graphics  
We present a semi-supervised co-analysis method for learning 3D shape styles from projected feature lines, achieving style patch localization with only weak supervision.  ...  Given a collection of 3D shapes spanning multiple object categories and styles, we perform style co-analysis over projected feature lines of each 3D shape and then backproject the learned style features  ...  This work was supported in part by NSFC (61572507, 61532003, 61622212) for Kai Xu, NSERC (611370) and a gift grant from Adobe Research for Hao Zhang, and NSERC PDF for Ali Mahdavi-Amiri.  ... 
doi:10.1145/3182158 fatcat:gopfmqhstzairau3reccsm7qze

Fine-grained semi-supervised labeling of large shape collections

Qi-Xing Huang, Hao Su, Leonidas Guibas
2013 ACM Transactions on Graphics  
., shapes can belong to multiple classes) semi-supervised approach that takes as input a large shape collection of a given category with associated sparse and noisy labels, and outputs cleaned and complete  ...  Experimental results show that despite this variety, given very sparse and noisy initial labels, the new method yields results that are superior to state-of-the-art semi-supervised learning techniques.  ...  This work was supported by NSF grants FODAVA 808515 and CCF 1011228, AFOSR grant FA9550-12-1-0372, ONR MURI award N00014-13-1-0341, a Google Research Award, and the support of the Max Planck Center for  ... 
doi:10.1145/2508363.2508364 fatcat:fb2744fyuvginm6gorkcwdwwxe
« Previous Showing results 1 — 15 out of 32,893 results