ViVo: Visual Vocabulary Construction for Mining Biomedical Images

Bhattacharya, Arnab; Ljosa, Vebjorn; Pan, Jia-Yu; Verardo, Mark R.; Yang, Hyung-Jeong; Faloutsos, Christos; Singh, Ambuj K.

doi:10.1109/icdm.2005.151

Cited by 21 publications

(15 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…It can detect clusters beyond Gaussians: clusters in full-dimensional data space as well as clusters in axisparallel subspaces (so called subspace-clusters) and in arbitrarily oriented subspaces (correlation clusters), and combinations and mixtures of clusters of all different types during one single run of the algorithm. 4. It can assign model distribution functions such as uniform, Gaussian, Laplacian (etc.)…”

Section: Contributionsmentioning

confidence: 99%

Robust information-theoretic clustering

Böhm¹,

Faloutsos²,

Pan³

et al. 2006

Proceedings of the 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

View full text Add to dashboard Cite

How do we find a natural clustering of a real world point set, which contains an unknown number of clusters with different shapes, and which may be contaminated by noise? Most clustering algorithms were designed with certain assumptions (Gaussianity), they often require the user to give input parameters, and they are sensitive to noise. In this paper, we propose a robust framework for determining a natural clustering of a given data set, based on the minimum description length (MDL) principle. The proposed framework, Robust Information-theoretic Clustering (RIC), is orthogonal to any known clustering algorithm: given a preliminary clustering, RIC purifies these clusters from noise, and adjusts the clusterings such that it simultaneously determines the most natural amount and shape (subspace) of the clusters. Our RIC method can be combined with any clustering technique ranging from K-means and K-medoids to advanced methods such as spectral clustering. In fact, RIC is even able to purify and improve an initial coarse clustering, even if we start with very simple methods such as grid-based space partitioning. Moreover, RIC scales well with the data set size. Extensive experiments on synthetic and real world data sets validate the proposed RIC framework. Figure 1: A fictitious dataset, (a) with a good clustering of one Gaussian cluster, one sub-space cluster, and noise; and (b) a bad clustering.

show abstract

Section: Contributionsmentioning

confidence: 99%

Robust information-theoretic clustering

Böhm¹,

Faloutsos²,

Pan³

et al. 2006

Proceedings of the 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

View full text Add to dashboard Cite

show abstract

“…However these low-level attributes usually cannot effectively capture semantics as previously mentioned. To overcome these problems some techniques use both low-level features in the form of visual keywords [6] and text annotation to perform content-based operations. However still the whole concept is static and does not take into consideration users' preferences etc.…”

Section: Introductionmentioning

confidence: 99%

On the Fly Semantic Annotation and Modelling of Multimedia

Doulamis

Ntalianis

2009

2009 16th International Conference on Systems, Signals and Image Processing

View full text Add to dashboard Cite

This paper introduces a novel framework for implicit media content annotation by putting the user in the loop of the annotation process. In particular, a ser of visual features are extracted to describe the visual properties of the media content. Then, as the user plays with the content by downloading a set of media data, a mechanisms associates the textual user's queries with the visual metadata. In this way, we achieve automatic media content annotation of the untagged media data by taking into account the "average" user's selection.

show abstract

“…The Visual Vocabulary (ViVo) [3] is a novel approach that uses Independent Component Analysis (ICA) to group image tiles into a set of visual terms, avoiding subtle problems (such as nonGaussianity) which hurt other clustering and dimensionality reduction methods. It was developed for use with classification of biomedical images.…”

Section: Clusteringmentioning

confidence: 99%

“…[5] [11] In biology, given a collection of fly embryos [15] or protein localization patterns [10] or cat retina images [3] and their labels, we want a system to answer the same types of questions.…”

Section: Introductionmentioning

confidence: 99%

QMAS: Querying, Mining and Summarization of Multi-modal Databases

Cordeiro

Guo

Haverkamp³

et al. 2010

2010 IEEE International Conference on Data Mining

View full text Add to dashboard Cite

Given a large collection of images, very few of which have labels given a priori, how can we automatically assign the labels of the remaining majority, and make suggestion for images that may need brand new labels distinct from existing ones? Popular automatic labeling techniques usually scale super linearly with the size of the image set, and/or their performances degrade if limited images bear initial labels. In this paper, we propose QMAS, an efficient solution to the following problems: (i) low-labor labeling (L3) -given a collection of images, very few of which are already labeled with keywords, find the most suitable labels for the remaining ones; and (ii) mining and attention routing -with the same input set, output a number of top representative images and top outliers. We present experimental evaluation on three data sets of proprietary and public satellite images up to a size of 2.25GB. QMAS scales linearly with the number of images, obtaining better or equal accuracy while being up to 40 times faster than its baseline algorithm. With limited numbers of initial labels available, QMAS achieves a significant accuracy margin over the baseline approach. The application of QMAS to recommend representatives and spot outliers is also illustrated. The proposed framework could be generalized to solve similar content-based annotation and mining problems on other multi-modal databases.

show abstract

ViVo: Visual Vocabulary Construction for Mining Biomedical Images

Cited by 21 publications

References 17 publications

Robust information-theoretic clustering

Robust information-theoretic clustering

On the Fly Semantic Annotation and Modelling of Multimedia

QMAS: Querying, Mining and Summarization of Multi-modal Databases

Contact Info

Product

Resources

About