Filters








3,492 Hits in 7.6 sec

Active Self-Paced Learning for Cost-Effective and Progressive Face Identification

Liang Lin, Keze Wang, Deyu Meng, Wangmeng Zuo, Lei Zhang
2018 IEEE Transactions on Pattern Analysis and Machine Intelligence  
By naturally combining two recently rising techniques: active learning (AL) and self-paced learning (SPL), our framework is capable of automatically annotating new instances and incorporating them into  ...  This paper aims to develop a novel cost-effective framework for face identification, which progressively maintains a batch of classifiers with the increasing face images of different individuals.  ...  As for the commercial product, the website "Face.com" once provided an API (application interface) to automatically detect and recognize faces in photos.  ... 
doi:10.1109/tpami.2017.2652459 pmid:28092522 fatcat:wvcef7tsx5djfmxq6t25ms4rf4

Complex Emotion Profiling: An Incremental Active Learning based Approach with Sparse Annotations

Selvarajah Thuseethan, Sutharshan Rajasegarar, John Yearwood
2020 IEEE Access  
In this paper, we propose a deep framework to incrementally and actively profile in-the-wild complex emotions, from sparse data.  ...  Therefore, it is important to learn the profile of the in-the-wild complex emotions accurately using limited annotated samples.  ...  The intuition behind this algorithm is to select the samples with high confidence values to automatically annotate and add them into the training set.  ... 
doi:10.1109/access.2020.3015917 fatcat:xbtqi423nfaolkk3vturgpaqpe

Face Recognition via Active Annotation and Learning

Hao Ye, Weiyuan Shao, Hong Wang, Jianqi Ma, Li Wang, Yingbin Zheng, Xiangyang Xue
2016 Proceedings of the 2016 ACM on Multimedia Conference - MM '16  
In this paper, we introduce an active annotation and learning framework for the face recognition task.  ...  Starting with an initial label deficient face image training set, we iteratively train a deep neural network and use this model to choose the examples for further manual annotation.  ...  Acknowledgment This work was supported in part by grants from Natural Science Foundation of China (No. 61572138) and Science and Technology Commission of Shanghai Municipality (No. 16511104802).  ... 
doi:10.1145/2964284.2984059 dblp:conf/mm/YeSWMWZX16 fatcat:jmjzpsnxnbcd3lcyilayfvrnxe

Learn the New, Keep the Old: Extending Pretrained Models with New Anatomy and Images [chapter]

Firat Ozdemir, Philipp Fuernstahl, Orcun Goksel
2018 Lecture Notes in Computer Science  
With ever increasing amounts of annotated medical datasets, it is infeasible to train a learning method always with all data from scratch.  ...  In this paper, we introduce a framework for applying incremental learning for segmentation and propose novel methods for selecting representative data therein.  ...  sample is represented with a class-discriminating vector.  ... 
doi:10.1007/978-3-030-00937-3_42 fatcat:3wyfgbqf7fctzko5tqztsvgf7e

Cost-Effective Active Learning for Deep Image Classification

Keze Wang, Dongyu Zhang, Ya Li, Ruimao Zhang, Liang Lin
2017 IEEE transactions on circuits and systems for video technology (Print)  
in an incremental learning manner.  ...  Specifically, these high confidence samples are automatically selected and iteratively assigned pseudo-labels.  ...  Thus it has great practical significance to develop a framework by combining CNNs and active learning, which can jointly learn features and classifiers/models from unlabeled training data with minimal  ... 
doi:10.1109/tcsvt.2016.2589879 fatcat:6vc55x3cjbhyxb4jcntirob57y

Self-Learning Scene-Specific Pedestrian Detectors Using a Progressive Latent Model

Qixiang Ye, Tianliang Zhang, Wei Ke, Qiang Qiu, Jie Chen, Guillermo Sapiro, Baochang Zhang
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Compared with conventional latent models, the proposed PLM incorporates a spatial regularization term to reduce ambiguities in object proposals and to enforce object localization, and also a graph-based  ...  Extensive experiments demonstrate that even without annotation the proposed self-learning approach outperforms weakly supervised learning approaches, while achieving comparable performance with transfer  ...  Tekes and Infotech Oulu are gratefully acknowledged.  ... 
doi:10.1109/cvpr.2017.222 dblp:conf/cvpr/YeZKQCSZ17 fatcat:4b6uc7qvfvhkhj5gljgim5a7ai

Field Extraction from Administrative Documents by Incremental Structural Templates

Marcal Rusinol, Tayeb Benkhelfallah, Vincent Poulain dAndecy
2013 2013 12th International Conference on Document Analysis and Recognition  
In this paper we present an incremental framework aimed at extracting field information from administrative document images in the context of a Digital Mail-room scenario.  ...  This model is incrementally refined as the system processes more and more documents from the same class.  ...  Finally, if the confidence of the extracted field is high enough, an incremental step is applied in order to adjust the learned structural model.  ... 
doi:10.1109/icdar.2013.223 dblp:conf/icdar/RusinolBD13 fatcat:dxa5wu337rd4nfoi7emy4pmcfa

A Weakly Supervised Method for Mud Detection in Ores Based on Deep Active Learning

Zhijian Huang, Fangmin Li, Xidao Luan, Zuowei Cai
2020 Mathematical Problems in Engineering  
Automatically detecting mud in bauxite ores is important and valuable, with which we can improve productivity and reduce pollution.  ...  The method uses the YOLO-v3 model as the basic detector, which is initialized with the pretrained weights on the MS COCO dataset. Then, an AL framework-embedded YOLO-v3 model is constructed.  ...  Acknowledgments is study was supported by the Scientific Research Fund of Hunan Provincial Education Department (nos. 18A376 and XJK17BXX010) and National Natural Science Foundation of China (no. 11701172  ... 
doi:10.1155/2020/3510313 fatcat:6klk3wmobbfr3ni3l4e6gtxuvm

Unsupervised Domain Adaptation in Semantic Segmentation: a Review [article]

Marco Toldo, Andrea Maracani, Umberto Michieli, Pietro Zanuttigh
2020 arXiv   pre-print
This problem has been recently explored and has rapidly grown with a large number of ad-hoc approaches.  ...  analysis of the classifier discrepancies, self-teaching, entropy minimization, curriculum learning and multi-task learning.  ...  Target segmentation maps are then directly filtered according to some confidence-based thresholding policy and used in combination with original source annotated data for the supervised learning of the  ... 
arXiv:2005.10876v1 fatcat:7t5v6qibxnfcxhwtohqqunhd2u

Personalized Image Classification by Semantic Embedding and Active Learning

Mofei Song
2020 Entropy  
To satisfactorily personalize the requirement, we propose an interactive image classification system with an offline representation learning stage and an online classification stage.  ...  To provide a high interactive rate, a unified active learning algorithm is used to search the optimal annotation and verification set by minimizing the expected time cost.  ...  [19] propose a cost-effective sample selection method to choose many high-confidence samples for deep learning.  ... 
doi:10.3390/e22111314 pmid:33287081 fatcat:ng2ablbtarfk7kyc27eclzrqri

Self-learning Scene-specific Pedestrian Detectors using a Progressive Latent Model [article]

Qixiang Ye, Tianliang Zhang, Qiang Qiu, Baochang Zhang, Jie Chen, Guillermo Sapiro
2016 arXiv   pre-print
Compared with conventional latent models, the proposed PLM incorporates a spatial regularization term to reduce ambiguities in object proposals and to enforce object localization, and also a graph-based  ...  Extensive experiments demonstrate that even without annotation the proposed self-learning approach outperforms weakly supervised learning approaches, while achieving comparable performance with transfer  ...  Acknowledgement The partial support of this work by ONR, NGA, ARO, NSF, NSFC under Grant 61271433 and 61671427, and Beijing Municipal Science & Technology Commission under Grant Z161100001616005 is gratefully  ... 
arXiv:1611.07544v1 fatcat:d6fvdn2hjzgurmd2nquvcli7mu

A Semi-Automatic Annotation Approach for Human Activity Recognition

Patrícia Bota, Joana Silva, Duarte Folgado, Hugo Gamboa
2019 Sensors  
Our approach uses a criterion to select the most relevant samples for annotation by the expert and propagate their label to the most confident samples.  ...  We introduce a Semi-Supervised Active Learning (SSAL) based on Self-Training (ST) approach for Human Activity Recognition to partially automate the annotation process, reducing the annotation effort and  ...  ST-SSAL displays during the initial iterations an automatic annotation percentage similar to AL and PL, with only the expert annotator labelling new samples and no automatic annotation, since the 0.98  ... 
doi:10.3390/s19030501 fatcat:svws6m4uovfvvkw23asebqxumy

Unsupervised Domain Adaptation in Semantic Segmentation: A Review

Marco Toldo, Andrea Maracani, Umberto Michieli, Pietro Zanuttigh
2020 Technologies  
This field has been recently explored and has rapidly grown with a large number of ad-hoc approaches.  ...  analysis of the classifier discrepancies, self-teaching, entropy minimization, curriculum learning and multi-task learning.  ...  Target segmentation maps are then directly filtered according to some confidence-based thresholding policy and used in combination with original source annotated data for the supervised learning of the  ... 
doi:10.3390/technologies8020035 fatcat:qzgjjiw5p5bldk76mh3s3pwlfq

Learning Semantic Part-Based Models from Google Images

Davide Modolo, Vittorio Ferrari
2018 IEEE Transactions on Pattern Analysis and Machine Intelligence  
Our framework works incrementally, by learning from easy examples first, and then gradually adapting to harder ones.  ...  We learn these rich models by collecting training instances for both parts and objects, and automatically connecting the two levels.  ...  Our framework instead does not require annotations of part positions nor extent, and automatically learns from Google Images instead. Learning from image search engines.  ... 
doi:10.1109/tpami.2017.2724029 pmid:28692966 fatcat:nibrsnahsfgxffyqj2ozrpgn5i

Visual Active Learning for Labeling: A Case for Soundscape Ecology Data

Liz Huancapaza Hilasaca, Milton Cezar Ribeiro, Rosane Minghim
2021 Information  
In this paper, we report the development of a methodology and a framework to support labeling, with an application case as background.  ...  The methodology performs visual active learning and label propagation with 2D embeddings as layouts to achieve faster and interactive labeling of samples.  ...  Thus, instances with low confidence scores are delivered to expert users to be labeled and instances with high confidence scores are used in the prediction automatically.  ... 
doi:10.3390/info12070265 fatcat:yvs2a5rdzfdifi3c5gsuozpl3q
« Previous Showing results 1 — 15 out of 3,492 results