Filters








2,344 Hits in 5.5 sec

Predicting memorability of images using attention-driven spatial pooling and image semantics

Bora Celikkale, Aykut Erdem, Erkut Erdem
2015 Image and Vision Computing  
In particular, we present an attention-driven spatial pooling strategy and show that considering image features from the salient parts of images improves the results of the previous models.  ...  Our goal in this article is to explore the role of visual attention and image semantics in understanding image memorability.  ...  Acknowledgments This work was supported by a grant from The Scientific and Technological Research Council of Turkey (TUBITAK) -Career Development Award 112E146.  ... 
doi:10.1016/j.imavis.2015.07.001 fatcat:qdbb4yfplberjbcchcqlukh22m

Visual Attention-Driven Spatial Pooling for Image Memorability

Bora Celikkale, Aykut Erdem, Erkut Erdem
2013 2013 IEEE Conference on Computer Vision and Pattern Recognition Workshops  
In particular, we present an attention-driven spatial pooling strategy for image memorability and show that the regions estimated by bottom-up and object-level saliency maps are more effective in predicting  ...  In daily life, humans demonstrate astounding ability to remember images they see on magazines, commercials, TV, the web and so on, but automatic prediction of intrinsic memorability of images using computer  ...  Acknowledgments This research was supported in part by The Scientic and Technological Research Council of Turkey (TUBITAK), Career Development Award 112E146.  ... 
doi:10.1109/cvprw.2013.142 dblp:conf/cvpr/CelikkaleEE13 fatcat:4y6vzrjk4bemhdxghickd5sw5u

Memorability of Image Regions

Aditya Khosla, Jianxiong Xiao, Antonio Torralba, Aude Oliva
2012 Neural Information Processing Systems  
Recent works have shown that image memorability is an intrinsic property of an image that can be reliably estimated using state-of-the-art image features and machine learning algorithms.  ...  In this work, we propose a probabilistic framework that models how and which local regions from an image may be forgotten using a data-driven approach that combines local and global images features.  ...  Acknowledgements We thank Phillip Isola and the reviewers for helpful discussions.  ... 
dblp:conf/nips/KhoslaXTO12 fatcat:3ih73tx5m5ew7giew7x4hqm75e

Dublin's Participation in the Predicting Media Memorability Task at MediaEval 2018

Alan F. Smeaton, Owen Corrigan, Paul Dockree, Cathal Gurrin, Graham Healy, Feiyan Hu, Kevin McGuinness, Eva Mohedano, Tomás Ward
2018 MediaEval Benchmarking Initiative for Multimedia Evaluation  
This paper outlines 6 approaches taken to computing video memorability, for the MediaEval Predicting Media Memorability Task.  ...  The approaches are based on video features, an end-to-end approach, saliency, aesthetics, neural feedback, and an ensemble of all approaches.  ...  Using Video and Image Saliency Visual saliency models generate a probability map highlighting image regions that most attract human attention.  ... 
dblp:conf/mediaeval/SmeatonCDGHHMMW18 fatcat:jfrsfcxlczf6nguqhthkm4aw6e

Memorability: An image-computable measure of information utility [article]

Zoya Bylinskii, Lore Goetschalckx, Anelise Newman, Aude Oliva
2021 arXiv   pre-print
In this chapter, we zoom into memorability with a computational lens, detailing the state-of-the-art algorithms that accurately predict image memorability relative to human behavioral data, using image  ...  Observer independence is what makes memorability an image-computable measure of information, and eligible for automatic prediction.  ...  They report that spatially pooling and weighting image features (e.g., GIST, SIFT, HOG) by the VMS not only yields better SVR predictions than without doing spatial pooling, but also outperforms spatial  ... 
arXiv:2104.00805v1 fatcat:ckncvcnyavedxkbxyfmu6gon5a

Shared memories driven by the intrinsic memorability of items [article]

Wilma A. Bainbridge
2021 arXiv   pre-print
Certain items -- including certain faces, words, images, and movements -- are intrinsically memorable or forgettable across observers, regardless of individual differences.  ...  In this chapter, I will discuss our current state-of-the-art understanding of memorability for visual information, and what these findings imply about how we perceive and remember visual events.  ...  Image color is not highly predictive of image memorability (Isola et al., 2011) , nor is spatial frequency .  ... 
arXiv:2104.06937v1 fatcat:infxy6tjfjbjtjbdr7dturz6oy

Predicting Image Memorability by Multi-view Adaptive Regression

Houwen Peng, Kai Li, Bing Li, Haibin Ling, Weihua Xiong, Weiming Hu
2015 Proceedings of the 23rd ACM international conference on Multimedia - MM '15  
Experimental results on the MIT benchmark dataset show the superiority of the proposed model compared with existing image memorability prediction methods.  ...  This phenomenon is caused by the intrinsic memorability of images revealed by recent studies [5, 6] .  ...  [2] introduce an attention-driven spatial pooling strategy for feature encoding and adopt SVR for prediction.  ... 
doi:10.1145/2733373.2806303 dblp:conf/mm/PengLLLXH15 fatcat:dcqy34c5t5aglcvzyeiewk5ave

LSTM-CF: Unifying Context Modeling and Fusion with LSTMs for RGB-D Scene Labeling [article]

Zhen Li, Yukang Gan, Xiaodan Liang, Yizhou Yu, Hui Cheng, Liang Lin
2016 arXiv   pre-print
and long-range spatial dependencies in an image along the vertical direction.  ...  Semantic labeling of RGB-D scenes is crucial to many intelligent applications including perceptual robotics.  ...  In future, we will explore how to extend the memorized layers with an attention mechanism, and refine the performance of our model in boundary labeling. Fig. 6 .  ... 
arXiv:1604.05000v3 fatcat:hs33ywcpbraxhieubxv3odt23i

LSTM-CF: Unifying Context Modeling and Fusion with LSTMs for RGB-D Scene Labeling [chapter]

Zhen Li, Yukang Gan, Xiaodan Liang, Yizhou Yu, Hui Cheng, Liang Lin
2016 Lecture Notes in Computer Science  
and longrange spatial dependencies in an image along the vertical direction.  ...  Semantic labeling of RGB-D scenes is crucial to many intelligent applications including perceptual robotics.  ...  In future, we will explore how to extend the memorized layers with an attention mechanism, and refine the performance of our model in boundary labeling. Fig. 6 .  ... 
doi:10.1007/978-3-319-46475-6_34 fatcat:lofxpg2obfenrjufd55vwcrdy4

Simultaneous Mapping and Target Driven Navigation [article]

Georgios Georgakis, Yimeng Li, Jana Kosecka
2019 arXiv   pre-print
We demonstrate that the use of semantic information improves localization accuracy and the ability of storing spatial semantic map aids the target driven navigation policy.  ...  The semantic and appearance stored in 2.5D map is distilled from RGB images, semantic segmentation and outputs of object detectors by convolutional neural networks.  ...  The mapping component leverages the outputs of object detection and semantic segmentation to construct a spatial representation of a scene which contains some semantic information.  ... 
arXiv:1911.07980v1 fatcat:n2onyhoebzapri7bx7qjjwzo5y

Learning deep representations for semantic image parsing: a comprehensive overview

Lili Huang, Jiefeng Peng, Ruimao Zhang, Guanbin Li, Liang Lin
2018 Frontiers of Computer Science  
Semantic image parsing, which refers to the process of decomposing images into semantic regions and constructing the structure representation of the input, has recently aroused widespread interest in the  ...  In this paper, we summarize three aspects of the progress of research on semantic image parsing, i.e., category-level semantic segmentation, instance-level semantic segmentation, and beyond segmentation  ...  Works beyond segmentation not only semantically segment images but also predict richer and finer results, such as the structures and relations of objects and the spatial layout.  ... 
doi:10.1007/s11704-018-7195-8 fatcat:p5hvfwhl5rbork5vf4rpnx3h6u

Relation Network for Multilabel Aerial Image Classification

Yuansheng Hua, Lichao Mou, Xiao Xiang Zhu
2020 IEEE Transactions on Geoscience and Remote Sensing  
The label relational inference module finally predicts label existences using label relations reasoned from outputs of the previous module.  ...  Multilabel classification plays a momentous role in perceiving intricate contents of an aerial image and triggers several related studies over the last years.  ...  However, spatial contextual semantics are not taken into account in this way. To address such an issue, here, we make use of 1 × 1 convolution instead of an MLP to explore spatial information.  ... 
doi:10.1109/tgrs.2019.2963364 fatcat:lpk2hmjhc5doplnxuyvyyen3fy

ResMem-Net: memory based deep CNN for image memorability estimation

Arockia Praveen, Abdulfattah Noorwali, Duraimurugan Samiayya, Mohammad Zubair Khan, Durai Raj Vincent P M, Ali Kashif Bashir, Vinoth Alagupandi
2021 PeerJ Computer Science  
But due to the introduction of Deep Learning and the large availability of data and GPUs, great strides have been made in predicting the memorability of an image.  ...  We have also used the heatmaps and results to analyze and answer one of the most important questions in image memorability: "What makes an image memorable?".  ...  The idea of using data-driven strategies to predict image memorability was first introduced by Isola et al. (2011a) .  ... 
doi:10.7717/peerj-cs.767 pmid:34825056 pmcid:PMC8594589 fatcat:xptm6a3fj5aevknxc547qcco4i

Coastal Land Cover Classification of High-Resolution Remote Sensing Images Using Attention-Driven Context Encoding Network

Jifa Chen, Gang Chen, Lizhe Wang, Bo Fang, Ping Zhou, Mingjie Zhu
2020 Sensors  
Therefore, this paper proposes a novel attention-driven context encoding network to solve these problems.  ...  sensing images.  ...  funding this study; Fengliang Fan for providing the datasets and hardware support used in this study.  ... 
doi:10.3390/s20247032 pmid:33302547 pmcid:PMC7763023 fatcat:p7wpwg4mxna4dj6gq7bckbauqu

Relation Network for Multi-label Aerial Image Classification [article]

Yuansheng Hua, Lichao Mou, Xiao Xiang Zhu
2020 arXiv   pre-print
The label relational inference module finally predicts label existences using label relations reasoned from outputs of the previous module.  ...  Multi-label classification plays a momentous role in perceiving intricate contents of an aerial image and triggers several related studies over the last years.  ...  However, spatial contextual semantics are not taken into account in this way. To address such issue, here, we make use of 1 × 1 convolution instead of an MLP to explore spatial information.  ... 
arXiv:1907.07274v3 fatcat:2lqaw6qxp5cavel44qytlz7ali
« Previous Showing results 1 — 15 out of 2,344 results