Filters








5,644 Hits in 5.9 sec

Unsupervised Learning of Semantics of Object Detections for Scene Categorization [chapter]

Grégoire Mesnil, Salah Rifai, Antoine Bordes, Xavier Glorot, Yoshua Bengio, Pascal Vincent
2014 Advances in Intelligent Systems and Computing  
In this paper, we propose to study different approaches to efficiently learn contextual semantics out of these object detections.  ...  for categorizing scenes.  ...  Codes for the experiments have been implemented using Theano [4] Machine Learning library.  ... 
doi:10.1007/978-3-319-12610-4_13 fatcat:a36rlagtizg4levkrxozjaiehq

A Survey of Visual Sensory Anomaly Detection [article]

Xi Jiang, Guoyang Xie, Jinbao Wang, Yong Liu, Chengjie Wang, Feng Zheng, Yaochu Jin
2022 arXiv   pre-print
Compared with semantic anomaly detection which detects anomaly at the label level (semantic shift), visual sensory AD detects the abnormal part of the sample (covariate shift).  ...  Visual sensory anomaly detection (AD) is an essential problem in computer vision, which is gaining momentum recently thanks to the development of AI for good.  ...  For example, in indoor scenes, there will be more occlusion, so how to detect novel objects in more and more complex scenes remains a challenge. 4 Event-level AD Unsupervised Method The mainstream of  ... 
arXiv:2202.07006v1 fatcat:2bqzmmrnjzggti5tcewa3mh3sa

Hierarchical Semantics of Objects (hSOs)

Devi Parikh, Tsuhan Chen
2007 2007 IEEE 11th International Conference on Computer Vision  
for enhanced object detection.  ...  The proposed approach, being entirely unsupervised, can detect the parts of the images that belong to the foreground objects, cluster these parts to represent objects, and provide an understanding of the  ...  Acknowledgments We thank Andrew Stein and Dhruv Batra for code to compute geometrically compatible correspondences.  ... 
doi:10.1109/iccv.2007.4408960 dblp:conf/iccv/ParikhC07 fatcat:k4dpypjnl5btzjazxgokdmjtni

FAN: Focused Attention Networks [article]

Chu Wang, Babak Samari, Vladimir Kim, Siddhartha Chaudhuri, Kaleem Siddiqi
2019 arXiv   pre-print
Thus far the learning of attention weights has been driven solely by the minimization of task specific loss functions.  ...  We further introduce a focused attention backbone to learn these attention weights for general tasks.  ...  It not only boosts performance in object detection, scene categorization and document classification, but also leads to state-of-the-art performance in a relationship proposal task.  ... 
arXiv:1905.11498v3 fatcat:z723v3lddba63cu3ze5a2nixz4

TV ad video categorization with probabilistic latent concept learning

Jinqiao Wang, Lingyu Duan, Lei Xu, Hanqing Lu, Jesse S. Jin
2007 Proceedings of the international workshop on Workshop on multimedia information retrieval - MIR '07  
A bag-of-words representation is proposed to discover ad categories-related latent visual and textual concepts by probabilistic latent semantics analysis (PLSA).  ...  The digested ads can be considered the video-based alert for emerging products/services. Thus the reachability and focus of TV ads can be improved.  ...  This fact motivates us to consider an ad video as a bag of semantic meaningful concepts in terms of scenes or objects.  ... 
doi:10.1145/1290082.1290113 dblp:conf/mir/WangDXLJ07 fatcat:zls6oxtxybczxcehr5paz6yvmi

Unsupervised Learning of Hierarchical Semantics of Objects (hSOs)

Devi Parikh, Tsuhan Chen
2007 2007 IEEE Conference on Computer Vision and Pattern Recognition  
We propose an approach for the unsupervised learning of the hSO.  ...  In addition to providing the semantic layout of the scene, learnt hSOs can have several useful applications such as providing context for enhanced object detection and compact scene representation for  ...  Acknowledgments We thank Andrew Stein and Dhruv Batra for discussion and code to compute geometrically compatible correspondences among images.  ... 
doi:10.1109/cvpr.2007.383370 dblp:conf/cvpr/ParikhC07 fatcat:a5nmjn3dkbddzggdrkotmtrypu

Unsupervised Modeling of Objects and Their Hierarchical Contextual Interactions

Devi Parikh, Tsuhan Chen
2009 EURASIP Journal on Image and Video Processing  
We propose an approach for the unsupervised learning of the hSO from a collection of images of a particular scene.  ...  semantically meaningful grouping of these objects, which characterizes the entire scene.  ...  In this paper, we introduce hierarchical semantics of objects (hSOs). We propose an approach for unsupervised learning of hSO from a collection of images.  ... 
doi:10.1155/2009/184618 fatcat:zcziakwcpneenccc3x6ak4dvgy

Co-clustering for Auditory Scene Categorization

Rui Cai, Lie Lu, Alan Hanjalic
2008 IEEE transactions on multimedia  
In this paper, we focus on the problem of automatically categorizing audio scenes in unsupervised manner.  ...  Automatically classifying and grouping auditory scenes with similar semantics into categories is beneficial for many multimedia applications, such as semantic event detection and indexing.  ...  Many recent research works explored the role of auditory scenes for semantic event detection or indexing in audiovisual data (video) [2] - [6] .  ... 
doi:10.1109/tmm.2008.921739 fatcat:vox53lvwknhi5huwujb6jyyeju

Semantic indexing and computational aesthetics

Miriam Redi
2013 Proceedings of the 3rd ACM conference on International conference on multimedia retrieval - ICMR '13  
Semantic Indexing is about automatically identifying content in natural images, namely recognizing objects and scenes.  ...  We investigate the role of Semantic Indexing techniques for Computational Aesthetics Frameworks, and, vice versa, the importance of Aesthetic features for Semantic Indexing prediction.  ...  Semantic Indexing works generally by building frameworks for scene categorization using holistic features [8, 18] , for object recognition using local features [1] , or for concept detection for video  ... 
doi:10.1145/2461466.2461532 dblp:conf/mir/Redi13 fatcat:5tu5dqztkjcrrfgseyxcba63ve

Aesthetics in Digital Photography: A Survey

Madhura Phatak
2020 International Journal for Research in Applied Science and Engineering Technology  
Photo aesthetics is an emerging field of research which can categorize large image collections into aesthetically good and bad photographs.  ...  .Image aesthetics is a high level image retrieval attribute, we discuss and analyze the various aesthetics classification and prediction methods prevalent for determining the aesthetic inclination of  ...  In [10] the techniques for reducing the semantic gap are:-1) Object Ontology: Algorithms to applications with simple semantic features, for describing the high level concepts. 2) Machine Learning: This  ... 
doi:10.22214/ijraset.2020.2105 fatcat:voseaxud3rheliqpl4dckin33u

Semantics Extraction from Images [chapter]

Ioannis Pratikakis, Anastasia Bolovinou, Bassilios Gatos, Stavros Perantonis
2011 Lecture Notes in Computer Science  
An overview of the state-of-the-art on semantics extraction from images is presented.  ...  For each combination of knowledge and image representation, a detailed discussion is addressed that leads to fruitful conclusions for the impact of each approach.  ...  The proposed approach was tested for the detection of semantic objects on video data of three different domains. In [100] , an object categorization method is presented ( Figure 14 ).  ... 
doi:10.1007/978-3-642-20795-2_3 fatcat:h3xh2fmryfgj5hwn5k5fjsximy

Unsupervised Learning of Functional Categories in Video Scenes [chapter]

Matthew W. Turek, Anthony Hoogs, Roderic Collins
2010 Lecture Notes in Computer Science  
We view the problem from the perspective of categorical object recognition, and present an approach for unsupervised learning of functional scene element categories.  ...  Existing methods for video scene analysis are primarily concerned with learning motion patterns or models for anomaly detection.  ...  Here, we view the problem from the perspective of categorical object recognition, and present an approach for unsupervised learning and modeling of functional scene element categories -entities that are  ... 
doi:10.1007/978-3-642-15552-9_48 fatcat:ukvfmtzjanfo5baj4udqs6yesq

What makes instance discrimination good for transfer learning? [article]

Nanxuan Zhao and Zhirong Wu and Rynson W.H. Lau and Stephen Lin
2021 arXiv   pre-print
Notably, recent work on unsupervised pretraining has shown to surpass the supervised counterpart for finetuning downstream applications such as object detection and segmentation.  ...  From this understanding of instance discrimination, how can we better exploit human annotation labels for pretraining? Our findings are threefold.  ...  AN ANALYSIS FOR VISUAL TRANSFER LEARNING We study the transfer performance of pretrained models for a set of downstream tasks: object detection on PASCAL VOC07, object detection and instance segmentation  ... 
arXiv:2006.06606v2 fatcat:gyleg63lbzfqpbkb2b3aryz63u

A Discriminative Representation of Convolutional Features for Indoor Scene Recognition

Salman H. Khan, Munawar Hayat, Mohammed Bennamoun, Roberto Togneri, Ferdous A. Sohel
2016 IEEE Transactions on Image Processing  
To this end, we introduce a new large-scale dataset of 1300 object categories which are commonly present in indoor scenes.  ...  Traditionally used convolutional features preserve the global spatial structure, which is a desirable property for general object recognition.  ...  ACKNOWLEDGEMENTS This research was supported by the SIRF and IPRS scholarships from the University of Western Australia (UWA) and the Australian Research Council (ARC) grants DP110102166, DP150100294 and  ... 
doi:10.1109/tip.2016.2567076 pmid:28113718 fatcat:u2kt4yc55vb3rapc2eqyvyrssq

A Survey on Deep Learning for Localization and Mapping: Towards the Age of Spatial Machine Intelligence [article]

Changhao Chen, Bing Wang, Chris Xiaoxuan Lu, Niki Trigoni, Andrew Markham
2020 arXiv   pre-print
In this work, we provide a comprehensive survey, and propose a new taxonomy for localization and mapping using deep learning.  ...  It is our hope that this work can connect emerging works from robotics, computer vision and machine learning communities, and serve as a guide for future researchers to apply deep learning to tackle localization  ...  A good example is [109] that offers object-level semantic mapping by identifying individual objects via a bounding box detection module and an unsupervised geometric segmentation module.  ... 
arXiv:2006.12567v2 fatcat:snb2byqamfcblauw5lzccb7umy
« Previous Showing results 1 — 15 out of 5,644 results