Filters








23,634 Hits in 2.9 sec

Mapping visual features to semantic profiles for retrieval in medical imaging

Johannes Hofmanninger, Georg Langs
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
We show that these semantic profiles enable higher recall and precision during retrieval compared to visual features, and that we can even map semantic terms describing clinical findings from radiology  ...  To learn models that capture the relationship between semantic clinical information and image elements at scale, we have to rely on data generated during clinical routine (images and radiology reports)  ...  [20] : L l (f ) = 0 if (s sub l l ) · f sub l ≤ τ l 1 if (s sub l l ) · f sub l > τ l (8) For a split test, a set of dimensions from the feature space is randomly selected and only the corresponding  ... 
doi:10.1109/cvpr.2015.7298643 dblp:conf/cvpr/HofmanningerL15 fatcat:twhhvm7webeoroy2t2db67mtdi

UNIBA: Combining Distributional Semantic Models and Word Sense Disambiguation for Textual Similarity

Pierpaolo Basile, Annalina Caputo, Giovanni Semeraro
2014 Proceedings of the 8th International Workshop on Semantic Evaluation (SemEval 2014)  
The integration of similarity measures is performed by means of two supervised methods based on Gaussian Process and Support Vector Machine.  ...  We propose to combine the output of different semantic similarity measures which exploit Word Sense Disambiguation and Distributional Semantic Models, among other lexical features.  ...  to build the distributional space based on synsets (SDS) extracted from BabelNet (we used two cycles of retraining); • Weka for the supervised approach.  ... 
doi:10.3115/v1/s14-2133 dblp:conf/semeval/BasileCS14 fatcat:5r5bqc3ktzcgjcpj42bxbws4ym

A Semantic Image Category for Structuring TV Broadcast Video Streams [chapter]

Jinqiao Wang, Lingyu Duan, Hanqing Lu, Jesse S. Jin
2006 Lecture Notes in Computer Science  
In this paper, we propose a semantic image category, named as Program Oriented Informative Images (POIM), to facilitate the segmentation, indexing and retrieval of different programs.  ...  The recognition of POIM, together with other audiovisual features, can be used to further determine program boundaries.  ...  Aiming at effective indexing and retrieval, semantic concepts and ontologies have been proposed to bridge the semantic gap inherent to video content analysis.  ... 
doi:10.1007/11922162_33 fatcat:ayjra2qp5vhcnfp2fn7ndutjeq

Self-Supervised Class Incremental Learning [article]

Zixuan Ni, Siliang Tang, Yueting Zhuang
2021 arXiv   pre-print
To comprehensively discuss the difference in performance between supervised and self-supervised methods in CIL, we set up three different class incremental schemes: Random Class Scheme, Semantic Class  ...  Existing Class Incremental Learning (CIL) methods are based on a supervised classification framework sensitive to data labels.  ...  the samples in the same sub-dataset have similar semantic information while the samples in different sub-dataset only has a few overlapping in semantic space (such as animal, commodity, etc.).  ... 
arXiv:2111.11208v1 fatcat:dfhl4nu4prh6jhggmur4mfzsa4

Semi-supervised lung nodule retrieval [article]

Mark Loyman, Hayit Greenspan
2020 arXiv   pre-print
The current study suggests a semi-supervised approach that involves two steps: 1) Automatic annotation of a given partially labeled dataset; 2) Learning a semantic similarity metric space based on the  ...  The semi-supervised approach has demonstrated a significantly higher discriminative ability than the fully-unsupervised reference.  ...  The authors acknowledge the National Cancer Institute and the Foundation for the National Institutes of Health, and their critical role in the creation of the free publicly available LIDC/IDRI Database used  ... 
arXiv:2005.01805v1 fatcat:vm2ehvajfnarfmiekh6uu6jbwq

Weakly-Supervised Semantic Segmentation via Sub-category Exploration [article]

Yu-Ting Chang, Qiaosong Wang, Wei-Chih Hung, Robinson Piramuthu, Yi-Hsuan Tsai, Ming-Hsuan Yang
2020 arXiv   pre-print
Existing weakly-supervised semantic segmentation methods using image-level annotations typically rely on initial responses to locate object regions.  ...  To enforce the network to pay attention to other parts of an object, we propose a simple yet effective approach that introduces a self-supervised task by exploiting the sub-category information.  ...  With this method, the parent classification learns a feature space through supervised training via L p , while the sub-category objective L s explores the feature sub-space and provides additional gradients  ... 
arXiv:2008.01183v1 fatcat:gvz6csafdzerxg5hmpqvdowy2q

Weakly-Supervised Semantic Segmentation via Sub-Category Exploration

Yu-Ting Chang, Qiaosong Wang, Wei-Chih Hung, Robinson Piramuthu, Yi-Hsuan Tsai, Ming-Hsuan Yang
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Existing weakly-supervised semantic segmentation methods using image-level annotations typically rely on initial responses to locate object regions.  ...  To enforce the network to pay attention to other parts of an object, we propose a simple yet effective approach that introduces a self-supervised task by exploiting the sub-category information.  ...  With this method, the parent classification learns a feature space through supervised training via L p , while the sub-category objective L s explores the feature sub-space and provides additional gradients  ... 
doi:10.1109/cvpr42600.2020.00901 dblp:conf/cvpr/ChangWHPT020 fatcat:aearf7ahpvaldhpnq7kxyafbdi

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.  ...  We design a novel metric learning strategy that preserves semantic similarity between labeled data, and employ entropy regularization term to fully exploit inherent potentials of unlabeled data.  ...  Then, from the sub-vectors of the input feature vector, sub-binary code is obtained by replacing each sub-vector with the index of the nearest codeword in the codebook.  ... 
arXiv:2002.11281v3 fatcat:v27lks4kvffbrkviy7ubveuebi

Unbiased Subclass Regularization for Semi-Supervised Semantic Segmentation [article]

Dayan Guan, Jiaxing Huang, Aoran Xiao, Shijian Lu
2022 arXiv   pre-print
Semi-supervised semantic segmentation learns from small amounts of labelled images and large amounts of unlabelled images, which has witnessed impressive progress with the recent advance of deep neural  ...  We build the balanced subclass distributions by clustering pixels of each original class into multiple subclasses of similar sizes, which provide class-balanced pseudo supervision to regularize the class-biased  ...  To generate unbiased pseudo labels for the original class supervision, we first map the prediction p w u from the subclass space (1, C sub ) H×W to the original class space (1, C) H×W (this process denoted  ... 
arXiv:2203.10026v2 fatcat:fbmv64alxrcb7bwggg6efsqdqi

TaxoCom: Topic Taxonomy Completion with Hierarchical Discovery of Novel Topic Clusters [article]

Dongha Lee, Jiaming Shen, SeongKu Kang, Susik Yoon, Jiawei Han, Hwanjo Yu
2022 arXiv   pre-print
to be discriminative among known (i.e., given) sub-topics, and (ii) novelty adaptive clustering assigns terms into either one of the known sub-topics or novel sub-topics.  ...  within a hierarchical topic structure, TaxoCom devises its embedding and clustering techniques to be closely-linked with each other: (i) locally discriminative embedding optimizes the text embedding space  ...  Weakly supervised document classification using topic taxonomy.  ... 
arXiv:2201.06771v1 fatcat:kqhnz4a2vnb3ncg3mv2rqlddz4

UNet++: A Nested U-Net Architecture for Medical Image Segmentation [chapter]

Zongwei Zhou, Md Mahfuzur Rahman Siddiquee, Nima Tajbakhsh, Jianming Liang
2018 Lecture Notes in Computer Science  
The re-designed skip pathways aim at reducing the semantic gap between the feature maps of the encoder and decoder sub-networks.  ...  Our architecture is essentially a deeply-supervised encoder-decoder network where the encoder and decoder sub-networks are connected through a series of nested, dense skip pathways.  ...  These encoder-decoder networks used for segmentation share a key similarity: skip connections, which combine deep, semantic, coarse-grained feature maps from the decoder sub-network with shallow, low-level  ... 
doi:10.1007/978-3-030-00889-5_1 pmid:32613207 pmcid:PMC7329239 fatcat:tlcuw5okfngexos73kmhimkoya

Benchmarking Omni-Vision Representation through the Lens of Visual Realms [article]

Yuanhan Zhang, Zhenfei Yin, Jing Shao, Ziwei Liu
2022 arXiv   pre-print
Beyond pulling two instances from the same concept closer -- the typical supervised contrastive learning framework -- ReCo also pulls two instances from the same semantic realm closer, encoding the semantic  ...  Without semantic overlapping, these datasets cover most visual realms comprehensively and meanwhile efficiently.  ...  We illustrate three principles to select semantic realms from numerous sub-trees. Fig. 3 . 3 Fig.3. Supervised contrastive losses v.s ReCo.  ... 
arXiv:2207.07106v2 fatcat:rbqrhp47g5f3xhslqvotmidft4

Temporally Consistent Gaussian Random Field for Video Semantic Analysis

Jinhui Tang, Xian-Sheng Hua, Tao Mei, Guo-Jun Qi, Shipeng Li, Xiuqing Wu
2007 2007 IEEE International Conference on Image Processing  
However, for the application of video semantic annotation, these methods only consider the relations among samples in the feature space and neglect an intrinsic property of video data: the temporally adjacent  ...  video segments (e.g., shots) usually have similar semantic concept.  ...  Using TCGRF method, the 64256 sub-shots in EVAL set are labeled as f (subshot i ), and the sub-shots in the same shot are merged using the "max" rule: f (shot m ) = max subshoti∈shotm {f (subshot i )}  ... 
doi:10.1109/icip.2007.4380070 dblp:conf/icip/TangHMQLW07 fatcat:m3sa6fohqnhw7g4cxd5muhivuy

Image classification using hybrid neural networks

Chih-Fong Tsai, Ken McGarry, John Tait
2003 Proceedings of the 26th annual international ACM SIGIR conference on Research and development in informaion retrieval - SIGIR '03  
Use of semantic content is one of the major issues which needs to be addressed for improving image retrieval effectiveness.  ...  Content-based image retrieval, image indexing/classification, and neural networks.  ...  Semantic-Based Classifier After the number and names of semantic classes and training data are defined and selected using the SOM, a supervised neural network, Support Vector Machines (SVMs) [3] , is  ... 
doi:10.1145/860500.860536 fatcat:m3iosu5ie5cgvbv7lmkidmoj6q

Image classification using hybrid neural networks

Chih-Fong Tsai, Ken McGarry, John Tait
2003 Proceedings of the 26th annual international ACM SIGIR conference on Research and development in informaion retrieval - SIGIR '03  
Use of semantic content is one of the major issues which needs to be addressed for improving image retrieval effectiveness.  ...  Content-based image retrieval, image indexing/classification, and neural networks.  ...  Semantic-Based Classifier After the number and names of semantic classes and training data are defined and selected using the SOM, a supervised neural network, Support Vector Machines (SVMs) [3] , is  ... 
doi:10.1145/860435.860536 dblp:conf/sigir/TsaiMT03 fatcat:qypfsymp6jcmrhnuvvhulmf5ua
« Previous Showing results 1 — 15 out of 23,634 results