Image retrieval via probabilistic hypergraph ranking

Huang, Yuchi; Liu, Qingshan; Zhang, Shaoting; Metaxas, Dimitris N.

doi:10.1109/cvpr.2010.5540012

Cited by 240 publications

(139 citation statements)

References 27 publications

Supporting

Mentioning

139

Contrasting

Order By: Relevance

“…Huang et al [11] presented a Hypergraph Rank framework for retrieving images. In the weighted hypergraph images are vertices and image searching problem is considered as a problem of hypergraph ranking.…”

Section: B Hypergraph Based Methodsmentioning

confidence: 99%

Image Retrieval using Hypergraph of Visual Concepts

Kawale¹,

Kamalapur²

2017

IJITCS

View full text Add to dashboard Cite

Abstract-Retrieving images similar to query image from a large image collection is a challenging task. Image retrieval is most useful in the image search engine to find images similar to the query image. Most of the existing graph based image retrieval methods capture only pairwise similarity between images. The proposed work uses the hypergraph approach of the visual concepts. Each image can be represented by combination of the several visual concepts. Visual concept is the specific object or part of an image. There are several images in the database which can share multiple visual concepts. To capture such a relationship between group of images hypergraph is used. In proposed work, each image is considered as a vertex and each visual concept as a hyperedge in a hypergraph. All the images sharing same visual concept, form a hyperedge. Images in the dataset are represented using hypergraph. For each query image visual concept is identified. Similarity between query image and database image is identified. According to these similarities association scores are assigned to images, which will handle the image retrieval.

show abstract

Section: B Hypergraph Based Methodsmentioning

confidence: 99%

Image Retrieval using Hypergraph of Visual Concepts

Kawale¹,

Kamalapur²

2017

IJITCS

View full text Add to dashboard Cite

show abstract

“…The final ranking score vector is obtained by combining the independent label propagation (manifold ranking) results carried by each image in each view with different weights. -The averaged distance of multiple feature based metric (ADF) method [2], which constructs a single relevance graph using the metric of average distance from multiple views. -The unsupervised metric fusion (UMF) method [4], which conducts metric fusion without considering label propagation result.…”

Section: Methodsmentioning

confidence: 99%

“…In the semi-supervised metric fusion step, for each view we form a fused metric by combining the current metric of the view and the label propagation results from other views. Unlike the methods in [2,4] that obtain a fused metric from multi-views without label information, the metric fusion step in our method fully utilizes the label information from all views. In the label propagation step, in each view we conduct label propagation using the fused metric.…”

Section: Introductionmentioning

confidence: 99%

“…That is, they linearly and independently combine either the metrics or the label propagation (manifold ranking) results from multi-views and then build a model based on the combined results. Specifically, the metric fusion first strategy [2] obtains a linear fusion of the metrics from multi-views, constructs a single similarity graph based on the fused metric, and conducts label propagation or manifold ranking on the similarity graph. Alternatively, the propagation fusion strategy [3,5,6] conducts label propagation or manifold ranking on each view first, and then obtains a linear fusion of label propagation results based on label relevance ranking scores in multi-views as the overall label propagation results.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

An Iterative Fusion Approach to Graph-Based Semi-Supervised Learning from Multiple Views

Wang

Pei

Lin

et al. 2014

Advances in Knowledge Discovery and Data Mining

View full text Add to dashboard Cite

Abstract. Often, a data object described by many features can be naturally decomposed into multiple "views", where each view consists of a subset of features. For example, a video clip may have a video view and an audio view. Given a set of training data objects with multiple views, where some objects are labeled and the others are not, semi-supervised learning with graphs from multi-views tries to learn a classifier by treating each view as a similarity graph on all objects, where edges are defined by the similarity on object pairs based on the view attributes. Labels and label relevance ranking scores of labeled objects can be propagated from labeled objects to unlabeled objects on the similarity graphs so that similar objects receive similar labels. The state-of-the-art, onecombo-fits-all methods linearly and independently combine either the metrics or the label propagation results from multi-views and then build a model based on the combined results. However, more often than not, the similarities between various objects may be manifested differently by different views. In such situations, the one-combo-fits-all methods may not perform well. To tackle the problem, we develop an iterative SemiSupervised Metric Fusion (SSMF) approach in this paper. SSMF fuses metrics and label propagation results from multi-views iteratively until the fused metric and label propagation results converge simultaneously. Views are weighted dynamically during the fusion process so that the adversary effect of irrelevant views, identified at each iteration of fusion process, can be reduced effectively. To evaluate the effectiveness of SSMF, we apply it on multi-view based and content based image retrieval and multi-view based multi-label image classification on real world data set, which demonstrates that our method outperforms the state-of-the-art methods.

show abstract

“…Liu et.al proposed in [10], A transductive learning framework for image retrieval. It is based on a probabilistic Hypergraph.…”

Section: Fig1 Tagging Behavior Of Usermentioning

confidence: 99%