Latent visual context learning for web image applications

Zhou, Wengang; Tian, Qi; Lu, Yijuan; Yang, Linjun; Li, Houqiang

doi:10.1016/j.patcog.2010.08.016

Cited by 25 publications

(10 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The conventional quantization approach can be altered by using multi-vocabulary merging technique to achieve better performance [14]. We can also improve our current image similarity measure by considering discriminative capability of the inliers, such as IDF information or the weights of inliers suggested in [39], which is calculated based on the PageRank algorithm on visual word link graph.…”

Section: Discussionmentioning

confidence: 99%

“…For the same goal, Zhou et al proposed a multi-layer graph to rerank the image list based on their visual words, where PageRank approach was applied on the both visual word link graph and image link graph [39]. In [40], Deng et al suggested a co-regularized multi-graph learning framework, where the intra-graph and inter-graph constraints were imposed to achieve better retrieval performance by weak learning.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Finding more relevance: Propagating similarity on Markov random field for object retrieval

Peng

Zhu

et al. 2015

Signal Processing: Image Communication

View full text Add to dashboard Cite

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Finding more relevance: Propagating similarity on Markov random field for object retrieval

Peng

Zhu

et al. 2015

Signal Processing: Image Communication

View full text Add to dashboard Cite

“…Among these, query expansion [8][24] [25] reissues initial top-ranked results to find valuable features which are not present in the original query; spatial verification [26][5] filters those false-positives by checking geometric consistency of matched features, build efficient representation for false match filtering [7] [27], or extracts visual phrases [28] to verify matches on the more robust feature groups; and diffusion-based algorithms [29], including those based on graphs [9][30] [31] [32] propagate affinities or beliefs via a graph structure to capture the high-level connections between images. Also there are other methods aimed at selecting high-quality features [33], discovering feature co-occurrence [34] [35], extracting contextual information [36][37] [38], incorporating nearest-neighbor information [39] [40], adopting an alternative matching kernel [41], combining different searching results [42], or dealing with similarity between features [43]. Other recent proposed post-processing methods include [44] [45].…”

Section: E Post Processingmentioning

confidence: 99%

Heterogeneous Graph Propagation for Large-Scale Web Image Search

Xie

Zhou

Zhang

2015

IEEE Trans. on Image Process.

Self Cite

View full text Add to dashboard Cite

State-of-the-art web image search frameworks are often based on the bag-of-visual-words (BoVWs) model and the inverted index structure. Despite the simplicity, efficiency, and scalability, they often suffer from low precision and/or recall, due to the limited stability of local features and the considerable information loss on the quantization stage. To refine the quality of retrieved images, various postprocessing methods have been adopted after the initial search process. In this paper, we investigate the online querying process from a graph-based perspective. We introduce a heterogeneous graph model containing both image and feature nodes explicitly, and propose an efficient reranking approach consisting of two successive modules, i.e., incremental query expansion and image-feature voting, to improve the recall and precision, respectively. Compared with the conventional reranking algorithms, our method does not require using geometric information of visual words, therefore enjoys low consumptions of both time and memory. Moreover, our method is independent of the initial search process, and could cooperate with many BoVW-based image search pipelines, or adopted after other postprocessing algorithms. We evaluate our approach on large-scale image search tasks and verify its competitive search performance.

show abstract

“…c-f Saliency maps generated by LSP using patch decomposition in 8 × 8, 16 × 16, 32 × 32 and 64 × 64, respectively Similar to LGD, we decompose a source image into patches and initialize the saliency value of each patch p i,j with (1). Then we propagate the saliency from all the patches to p i,j , which is similar to VisualRank [22,50]:…”

Section: Location Based Saliency Detection Approachmentioning

confidence: 99%

How important is location information in saliency detection of natural images

Ren

Liu

et al. 2015

Multimed Tools Appl

View full text Add to dashboard Cite

Location information, i.e., the position of content in image plane, is considered as an important supplement in saliency detection. The effect of location information is usually evaluated by integrating it with the selected saliency detection methods and measuring the improvement, which is highly influenced by the selection of saliency methods. In this paper, we provide direct and quantitative analysis of the importance of location information for saliency detection in natural images. We firstly analyze the relationship between content location and saliency distribution on four public image datasets, and validate the distribution by simply treating location based Gaussian distribution as saliency map. To further validate the effectiveness of location information, we propose a location based saliency detection approach, which completely initializes saliency maps with location information and propagate saliency among patches based on color similarity, and discuss the robustness Multimed Tools Appl (2016) 75:2543-2564 of location information's effect. The experimental results show that location information plays a positive role in saliency detection, and the proposed method can outperform most state-of-the-art saliency detection methods and handle natural images with different object positions and multiple salient objects.

show abstract

Latent visual context learning for web image applications

Cited by 25 publications

References 30 publications

Finding more relevance: Propagating similarity on Markov random field for object retrieval

Finding more relevance: Propagating similarity on Markov random field for object retrieval

Heterogeneous Graph Propagation for Large-Scale Web Image Search

How important is location information in saliency detection of natural images

Contact Info

Product

Resources

About