Filters








2,666 Hits in 3.7 sec

Pairwise Similarity Knowledge Transfer for Weakly Supervised Object Localization [article]

Amir Rahimi, Amirreza Shaban, Thalaiyasingam Ajanthan, Richard Hartley, Byron Boots
2020 arXiv   pre-print
In this work, we argue that learning only an objectness function is a weak form of knowledge transfer and propose to learn a classwise pairwise similarity function that directly compares two input proposals  ...  For instance, in the ILSVRC dataset, the Correct Localization (CorLoc) performance improves from 72.8% to 78.2% which is a new state-of-the-art for WSOL task in the context of knowledge transfer.  ...  We gratefully express our gratitude to Judy Hoffman for her advice on improving the experiments. We also thank the anonymous reviewers for their helpful comments to improve the paper.  ... 
arXiv:2003.08375v2 fatcat:4jmfxg7krjcsni7rlpn2euvuhy

Weak Novel Categories without Tears: A Survey on Weak-Shot Learning [article]

Li Niu
2021 arXiv   pre-print
., noisy labels for image classification, image labels for object detection, bounding boxes for segmentation), similar to the definitions in weakly supervised learning.  ...  Therefore, weak-shot learning can also be treated as weakly supervised learning with auxiliary fully supervised categories.  ...  Despite the similarity in the sense of transferring knowledge from fully-annotated data to weakly-annotated data, weak-shot learning involves cross-category knowledge transfer, which is much more challenging  ... 
arXiv:2110.02651v2 fatcat:hvvniqgukrdktmu4vjj6fefuqq

Localizing Objects While Learning Their Appearance [chapter]

Thomas Deselaers, Bogdan Alexe, Vittorio Ferrari
2010 Lecture Notes in Computer Science  
Furthermore, our method enables to train any state-of-the-art object detector in a weakly supervised fashion, although it would normally require object location annotations.  ...  Our approach simultaneously localizes object instances while learning an appearance model specific for the class. We demonstrate this on the challenging Pascal VOC 2007 dataset.  ...  Initially Images with cars Images with localized cars Training a fully supervised detector Weakly Supervised Learning Weakly Supervised Learning Generic knowledge from other classes e.g.  ... 
doi:10.1007/978-3-642-15561-1_33 fatcat:ub6xww6bv5a5pfqs3nkmipjdwa

Weakly Supervised Localization and Learning with Generic Knowledge

Thomas Deselaers, Bogdan Alexe, Vittorio Ferrari
2012 International Journal of Computer Vision  
This shows that objectness is a powerful source of generic knowledge, which greatly helps localizing objects in weakly supervised images.  ...  weakly supervised learning of object classes.  ... 
doi:10.1007/s11263-012-0538-3 fatcat:agyvaorj3jh6thxu453njpp5fa

Learning to Find Common Objects Across Few Image Collections [article]

Amirreza Shaban, Amir Rahimi, Shray Bansal, Stephen Gould, Byron Boots, Richard Hartley
2019 arXiv   pre-print
Furthermore, we propose a fast greedy inference algorithm for energy minimization. We evaluate our approach on few-shot common object recognition as well as object co-localization tasks.  ...  Our experiments show that learning the pairwise and unary terms greatly improves the performance of the model over several well-known methods for these tasks.  ...  [32] propose a method to transfer knowledge from a set of familiar objects to localize new objects in a collection of weakly supervised images.  ... 
arXiv:1904.12936v2 fatcat:we4mbmlbfrhchihxddobk3borq

Weakly Supervised Object Co-Localization via Sharing Parts Based on a Joint Bayesian Model

Lu Wu, Quan Liu
2018 Symmetry  
To address this issue, we present a novel joint Bayesian model for weakly-supervised object localization.  ...  The differences compared to previous discriminative methods are evaluated in three aspects: (1) We co-localize the similar object per class through transferring shared parts, which are pooling by modeling  ...  Related Work Weakly Supervised Object Localization.  ... 
doi:10.3390/sym10050142 fatcat:uco5irnwnbb7hh5f3g5zqt4zcm

Mixed Supervised Object Detection by Transferring Mask Prior and Semantic Similarity [article]

Yan Liu, Zhijie Zhang, Li Niu, Junjie Chen, Liqing Zhang
2021 arXiv   pre-print
Moreover, the semantic similarity between objects learned from base categories is transferred to denoise the pseudo full annotations for novel categories.  ...  Previous works using mixed supervision mainly learn the class-agnostic objectness from fully-annotated categories, which can be transferred to upgrade the weak annotations to pseudo full annotations for  ...  The pairwise features are fed into SimNet and the predicted similarity scores A are supervised by the ground-truth similarities Â.  ... 
arXiv:2110.14191v1 fatcat:6g5bl2ni7nbpfpisyiglkbacym

Unsupervised Transfer Learning with Self-Supervised Remedy [article]

Jiabo Huang, Shaogang Gong
2020 arXiv   pre-print
Our method mitigates nontransferrable prior-knowledge by self-supervision, benefiting from both transfer and self-supervised learning.  ...  Extensive experiments on four datasets for image clustering tasks reveal the superiority of our model over the state-of-the-art transfer clustering techniques.  ...  There are other attempts at transferring knowledge across domains for achieving learning tasks that are similar to transfer clustering.  ... 
arXiv:2006.04737v1 fatcat:jivttxerg5chhaxo4qdwvtokiq

Weakly Supervised Dense Video Captioning via Jointly Usage of Knowledge Distillation and Cross-modal Matching [article]

Bofeng Wu, Guocheng Niu, Jun Yu, Xinyan Xiao, Jian Zhang, Hua Wu
2021 arXiv   pre-print
Extensive experiments on ActivityNet-Caption dataset reveal the significance of distillation-based event proposal generation and cross-modal retrieval-based semantic matching to weakly supervised DVC,  ...  This paper proposes an approach to Dense Video Captioning (DVC) without pairwise event-sentence annotation.  ...  teacher networks designed for similar tasks.  ... 
arXiv:2105.08252v1 fatcat:cgoqkex2szb6xbano66ivqkn24

Optimization Methods for Evaluating PEV Charging Considering Customer Behavior

Di Wu, Xinda Ke, Nikitha Radhakrishnan, Andrew Reiman
2018 2018 IEEE Power & Energy Society General Meeting (PESGM)  
To address this issue, we present a novel joint Bayesian model for weakly-supervised object localization.  ...  The differences compared to previous discriminative methods are evaluated in three aspects: (1) We co-localize the similar object per class through transferring shared parts, which are pooling by modeling  ...  Related Work Weakly Supervised Object Localization.  ... 
doi:10.1109/pesgm.2018.8586401 fatcat:fk2dywywpfhipegjwwaoqe2nl4

Unseen Object Segmentation in Videos via Transferable Representations [article]

Yi-Wen Chen, Yi-Hsuan Tsai, Chu-Ya Yang, Yen-Yu Lin, Ming-Hsuan Yang
2019 arXiv   pre-print
The entire process is decomposed into two tasks: 1) solving a submodular function for selecting object-like segments, and 2) learning a CNN model with a transferable module for adapting seen categories  ...  We present an iterative update scheme between two tasks to self-learn the final solution for object segmentation.  ...  In addition, our method is flexible for the weakly-supervised learning setting, which cannot be achieved by the above approaches. Object Segmentation in Weakly-supervised Videos.  ... 
arXiv:1901.02444v1 fatcat:bkg2fmito5dqdfuq7e7u6bgq4u

Integrating Object-aware and Interaction-aware Knowledge for Weakly Supervised Scene Graph Generation [article]

Xingchen Li, Long Chen, Wenbo Ma, Yi Yang, Jun Xiao
2022 arXiv   pre-print
Recently, increasing efforts have been focused on Weakly Supervised Scene Graph Generation (WSSGG).  ...  Hence, in this paper, we propose to enhance a simple grounding module with both object-aware and interaction-aware knowledge to acquire more reliable pseudo labels.  ...  the National Natural Science Foundation of China (U19B2043, 61976185), Zhejiang Natural Science Foundation (LR19F020002), Zhejiang Innovation Foundation (2019R52002), and the Fundamental Research Funds for  ... 
arXiv:2208.01834v1 fatcat:gbdmhnbwhjhovppeynfnhz7aaq

DiscoBox: Weakly Supervised Instance Segmentation and Semantic Correspondence from Box Supervision [article]

Shiyi Lan, Zhiding Yu, Christopher Choy, Subhashree Radhakrishnan, Guilin Liu, Yuke Zhu, Larry S. Davis, Anima Anandkumar
2021 arXiv   pre-print
Our best model achieves 37.9% AP on COCO instance segmentation, surpassing prior weakly supervised methods and is competitive to supervised methods.  ...  We also obtain state of the art weakly supervised results on PASCAL VOC12 and PF-PASCAL with real-time inference.  ...  During training, the model will retrieve pairwise images, we find all combinations of intra-class ob- similar intra-class objects from the object bank.  ... 
arXiv:2105.06464v2 fatcat:wr5iyiqvivb3novhhkgxwk6mv4

Deep Image Category Discovery using a Transferred Similarity Function [article]

Yen-Chang Hsu, Zhaoyang Lv, Zsolt Kira
2016 arXiv   pre-print
Such binary constraints can be learned from datasets in other domains as transferred similarity functions, which mimic a simple knowledge transfer.  ...  In this paper, we similarly utilize prior knowledge to facilitate the discovery of image categories.  ...  This ability in human learning motivates us to think about the following problem: how can we discover new object categories and structures in unlabeled data by weakly transferring knowledge from other  ... 
arXiv:1612.01253v1 fatcat:pfa6huoo6vgtlc35yc6y6uv5ay

AnchorNet: A Weakly Supervised Network to Learn Geometry-Sensitive Features for Semantic Matching

David Novotny, Diane Larlus, Andrea Vedaldi
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
of supervision is used.  ...  instances each from a different object class.  ...  They found that CNN features perform on par with hand-crafted alternatives such as SIFT for the weakly-supervised keypoint transfer problems, and can outperform them when keypoint supervision is available  ... 
doi:10.1109/cvpr.2017.306 dblp:conf/cvpr/NovotnyLV17 fatcat:4xivgkrkyjeqnpwxt3j3rthnc4
« Previous Showing results 1 — 15 out of 2,666 results