31,814 Hits in 5.3 sec

Improving Diversity in Image Search via Supervised Relevance Scoring

Eleftherios Spyromitros-Xioufis, Symeon Papadopoulos, Alexandru Lucian Ginsca, Adrian Popescu, Yiannis Kompatsiaris, Ioannis Vlahavas
2015 Proceedings of the 5th ACM on International Conference on Multimedia Retrieval - ICMR '15  
While relevance has drastically improved in recent years, diversity is still an open problem.  ...  Results returned by commercial image search engines should include relevant and diversified depictions of queries in order to ensure good coverage of users' information needs.  ...  performance. • We show that relevance scoring accuracy can be significantly improved by exploiting relevance annotations and performing the scoring in a supervised manner. • We evaluate supervised relevance  ... 
doi:10.1145/2671188.2749334 dblp:conf/mir/XioufisPGPKV15 fatcat:tuipupyennb35e2e4km36gehf4

Dense-Caption Matching and Frame-Selection Gating for Temporal Localization in VideoQA [article]

Hyounghun Kim, Zineng Tang, Mohit Bansal
2020 arXiv   pre-print
Dynamic spatio-temporal relationships between people/objects, and diverse multimodal events are present in a video clip.  ...  better supervise the model with human importance annotations.  ...  The views, opinions, and/or findings contained in this article are those of the authors and should not be interpreted as representing the official views or policies, either expressed or implied, of the  ... 
arXiv:2005.06409v1 fatcat:xvjyiheaevc5phpruyt4ixil2q

Tag relevance fusion for social image retrieval

Xirong Li
2014 Multimedia Systems  
We present a systematic study, covering tag relevance fusion in early and late stages, and in supervised and unsupervised settings.  ...  Experiments on a large present-day benchmark set show that tag relevance fusion leads to better image retrieval.  ...  propose to improve tag-based image search by visual-text joint hypergraph learning [7, 8] .  ... 
doi:10.1007/s00530-014-0430-9 fatcat:5z36n5pbjffmjc2aw3n3bamx5u

Search-based automatic image annotation via Flickr photos using tag expansion

Liang-Chi Hsieh, Winston H. Hsu
2010 2010 IEEE International Conference on Acoustics, Speech and Signal Processing  
To tackle such challenges, we propose a novel solution in three folds: (1) a tag expansion method to solve the sparsity of user-contributed tags; (2) improving tag relevance estimation by visual consistency  ...  In this work, we argue automatic image annotation by search over user-contributed photo sites (e.g., Flickr), which have accumulated rich human knowledge and billions of photos.  ...  Some previous works [1, 2] in (supervised) image categorization and classification are first proposed to annotate images.  ... 
doi:10.1109/icassp.2010.5496215 fatcat:6ysc67uoejgx3i4bdosuuoztlu

Image Relevance Prediction Using Query-Context Bag-of-Object Retrieval Model

Yang Yang, Linjun Yang, Gangshan Wu, Shipeng Li
2014 IEEE transactions on multimedia  
On the contrary, TOP_5_R only focus on image relevance but obtains low diversity score because there are too many duplicate images in the summarization.  ...  That is, to solve the image search reranking problem in a supervised way.  ... 
doi:10.1109/tmm.2014.2326836 fatcat:nrx3gx7kkrep5kwhceravor7iy

Image Search Reranking With Hierarchical Topic Awareness

Xinmei Tian, Linjun Yang, Yijuan Lu, Qi Tian, Dacheng Tao
2015 IEEE Transactions on Cybernetics  
TARerank describes the hierarchical topic structure of search results in one model, and seamlessly captures both relevance and diversity of the image search results simultaneously.  ...  Most of the traditional reranking methods cannot capture both relevance and diversity of the search results at the same time. Or they ignore the hierarchical topic structure of search result.  ...  BR improves the relevance of Text, while sacrificing diversity. Two-step method MMR-BR improves diversity and relevance in two separate steps and the errors are easily accumulated.  ... 
doi:10.1109/tcyb.2014.2366740 pmid:25616090 fatcat:btrfqvgbmfhbnhcouyt4ytv2xq

Image Tagging With Social Assistance

D. Dhayalan, M. Queen Mary Vidya, B. Gowri Priya
2018 International Journal of Trend in Scientific Research and Development  
In this paper, we advocate a topic diverse ranking method for tag-primarily based photo retrieval with the eye of selling the situation insurance overall performance.  ...  Tag-primarily based definitely image are seeking for is an essential method to find photos shared with the aid of manner of clients in social networks.  ...  Furthermore, a quantitative criterion called diversity scores (DS) is proposed to evaluate the improvement.  ... 
doi:10.31142/ijtsrd14566 fatcat:hejrlm66ajg4bdbthx642kor4a

Video search reranking via online ordinal reranking

Yi-Hsuan Yang, Winston H. Hsu
2008 2008 IEEE International Conference on Multimedia and Expo  
When evaluated in TRECVID search benchmark, ordinal reranking, while being extremely efficient, outperforms existing methods and offers 35.6% relative improvement over the text-based search baseline in  ...  The approach, ordinal reranking, reranks an initial search list by utilizing the co-occurrence patterns via the ranking functions such as ListNet.  ...  Reranking is then made by reordering the shots with high search scores linearly fused by these relevant concepts.  ... 
doi:10.1109/icme.2008.4607427 dblp:conf/icmcs/YangH08 fatcat:wjmbfhsr4zc4pevpxvlvdyqqmy

Joint Image Captioning and Question Answering [article]

Jialin Wu, Zeyuan Hu, Raymond J. Mooney
2018 arXiv   pre-print
Meanwhile, image captioning systems with beam search strategy tend to generate similar captions and fail to diversely describe images.  ...  supervision from answers.  ...  However, deep network systems still tend to generate similar captions with beam search strategy failing to diversely describe the images (Vijayakumar et al., 2016) .  ... 
arXiv:1805.08389v1 fatcat:cnjknisqcrd4pfyxgmioniuj2a

Query-adaptive Video Summarization via Quality-aware Relevance Estimation [article]

Arun Balajee Vasudevan, Michael Gygli, Anna Volokitin, Luc Van Gool
2017 arXiv   pre-print
Furthermore, we introduce a new dataset, annotated with diversity and query-specific relevance labels.  ...  Although the problem of automatic video summarization has recently received a lot of attention, the problem of creating a video summary that also highlights elements relevant to a search query has been  ...  ACKNOWLEDGEMENTS is work has been supported by Toyota via the project TRACE-Zurich. We also acknowledge the support by the CHIST-ERA project MUSTER.  ... 
arXiv:1705.00581v2 fatcat:w3sijnp4bnev3a4qng3n2oqqma

You Reap What You Sow: Using Videos to Generate High Precision Object Proposals for Weakly-Supervised Object Detection

Krishna Kumar Singh, Yong Jae Lee
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Our W-RPN proposals lead to significant improvement in performance for state-of-the-art weakly-supervised object detection approaches on PASCAL VOC 2007 and  ...  Existing weakly-supervised detection approaches use off-the-shelf proposal methods like edge boxes or selective search to obtain candidate boxes.  ...  This work was supported in part by NSF CAREER IIS-1751206, IIS-1748387, AWS ML Research Award, Google Cloud Platform research credits program, and GPUs donated by NVIDIA.  ... 
doi:10.1109/cvpr.2019.00964 dblp:conf/cvpr/SinghL19 fatcat:yecmw2pavvagfc6u4kh5mnvy5e

Query-adaptive Video Summarization via Quality-aware Relevance Estimation

Arun Balajee Vasudevan, Michael Gygli, Anna Volokitin, Luc Van Gool
2017 Proceedings of the 2017 ACM on Multimedia Conference - MM '17  
Furthermore, we introduce a new dataset, annotated with diversity and query-speci c relevance labels.  ...  Although the problem of automatic video summarization has recently received a lot of a ention, the problem of creating a video summary that also highlights elements relevant to a search query has been  ...  Clickture is a large dataset consisting of queries and retrieved images from Bing Image search. e annotation is in form of triplets (K, Q, C) meaning that the image K was clicked C times in the search  ... 
doi:10.1145/3123266.3123297 dblp:conf/mm/VasudevanGVG17 fatcat:lqe6pebxezc3bbvncz6wmiovla

Improving Generative Visual Dialog by Answering Diverse Questions

Vishvak Murahari, Prithvijit Chattopadhyay, Dhruv Batra, Devi Parikh, Abhishek Das
2019 Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)  
To improve this, we devise a simple auxiliary objective that incentivizes Q-BOT to ask diverse questions, thus reducing repetitions and in turn enabling A-BOT to explore a larger state space during RL  ...  fewer conflicting exchanges), fluent (i.e. more humanlike), and detailed, while still being comparably image-relevant as prior work and ablations.  ...  The Georgia Tech effort is supported in part by NSF, AFRL, DARPA, ONR YIPs, ARO PECASE. AD is supported in part by fellowships from Facebook, Adobe, and Snap Inc.  ... 
doi:10.18653/v1/d19-1152 dblp:conf/emnlp/MurahariCBPD19 fatcat:35dt2n4qa5govetclxcjw63kqe

Special issue on "visual semantic analysis with weak supervision"

Luming Zhang, Yang Yang, Rongrong Ji, Roger Zimmermann
2017 Multimedia Systems  
In "Graph-based Clustering and Ranking for Diversified Image Search", Yan et al. described a novel framework for Web image search results clustering and re-ranking, the goal is to improve diversity at  ...  In "Tag Relevance Fusion for Social Image Retrieval", Li et al. presented a systematic study, covering tag relevance fusion in the early and late stages, and in supervised and unsupervised settings.  ... 
doi:10.1007/s00530-016-0527-4 fatcat:72hcjiiwfzbk7mdrsumuia7rzy


Yi Hsuan Yang, Po Tun Wu, Ching Wei Lee, Kuan Hung Lin, Winston H. Hsu, Homer H. Chen
2008 Proceeding of the 16th ACM international conference on Multimedia - MM '08  
To tackle these issues, a novel search system, ContextSeer, is developed to improve search quality (by reranking) and recommend supplementary information (i.e., search-related tags and canonical images  ...  Second, to represent the diversity of search result, we propose an efficient algorithm cannoG to select multiple canonical images without clustering.  ...  in image/video search and navigation.  ... 
doi:10.1145/1459359.1459387 dblp:conf/mm/YangWLLHC08 fatcat:s7nthtjrxne4ndwzuebmcyu6cm
« Previous Showing results 1 — 15 out of 31,814 results