+ All Categories
Home > Data & Analytics > Metric s plat - a platform for quick development testing and visualization of content-based...

Metric s plat - a platform for quick development testing and visualization of content-based...

Date post: 08-Jun-2015
Category:
Upload: jose-f-rodrigues-jr
View: 34 times
Download: 1 times
Share this document with a friend
Description:
Jose Rodrigues, Luciana A S Romani, Luciana Zaina, Ricardo Ciferri (2009) MetricSPlat - A platform for quick development, testing and visualization of content-based retrieval techniques In: Simpósio Brasileiro de Bancos de Dados - SBBD2009 1-6. @inproceedings { RodriguesSBBD09, title = "MetricSPlat - A platform for quick development, testing and visualization of content-based retrieval techniques", year = "2009", author = "Jose Rodrigues and Luciana A S Romani and Luciana Zaina and Ricardo Ciferri", booktitle = "Simpósio Brasileiro de Bancos de Dados - SBBD2009", pages = "1-6", url = "http://www.icmc.usp.br/~junio/PublishedPapers/RodriguesSBBD09-MetricSPlat.pdf", urllink = "http://www.icmc.usp.br/~junio/MetricSPlat/index.htm", abstract = "The development and testing of content-based data retrieval systems is a time-consuming task. Over the concept of metric space, such systems must integrate the three factors that define an indexing environment. These factors are features extraction, metric structures and distance functions, not to mention a suitable user interface. This integration deviates the work from the real focus of research, suppressing quick experimentation of ideas. In this context, we present the Metric Space Platform (MetricSPlat), a system designed for content-based retrieval enabled with plug-in features. With minimal effort, MetricSPlat substantially speeds up the experimentation of new techniques by providing a well-defined framework aided with interactive data visualization techniques.", note = "8 pages", keywords = "visualization, content-based data retrieval"}
Popular Tags:
19
MetricSPlat - A platform for quick development, testing and visualization of content-based retrieval techniques Jose Fernando Rodrigues Junior – [email protected] Zaina, L. A. M. – [email protected] Romani, L. A. S. – [email protected] Ciferri, R. R. - [email protected] Universidade Federal de São Carlos – Campus Sorocaba Brasil Fortaleza - October/2009 http://www.icmc.usp.br/~junio/PublishedPapers/RodriguesSBBD09 -MetricSPlat.pdf
Transcript
Page 1: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

MetricSPlat - A platform for quick development, testing and visualization of content-based retrieval techniques

Jose Fernando Rodrigues Junior – [email protected]

Zaina, L. A. M. – [email protected] Romani, L. A. S. – [email protected]

Ciferri, R. R. - [email protected]

Universidade Federal de São Carlos – Campus SorocabaBrasil

Fortaleza - October/2009

http://www.icmc.usp.br/~junio/PublishedPapers/RodriguesSBBD09-MetricSPlat.pdf

Page 2: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Challenge:Challenge:

to ease the to ease the prototyping prototyping

and tuning of and tuning of CBIRCBIR

Page 3: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Metric spaces

Metric spaces: solution for non-orderable domains Images Videos Text Sound

It the case of imagesMetric spaces Content-based Image Retrieval (CBIR)

It demands three things: Features extraction A metric data structure (ideally) A distance function

Page 4: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Features Extraction

Metric spaces

Metric Structure Distance Function

Features Extraction

Page 5: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Features Extraction

Metric spaces

Metric Structure Distance Function

Features Extraction

Page 6: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Metric spaces

Putting together the three components of a metric space is a cumbersome task

The integration of these factors is systematic

However, it has been done ad hoc

Time is lost in re-sculpting the wheel

Page 7: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Contribution:Contribution:

a platform for a platform for

quick development quick development

and and

experimentation of experimentation of

CBIRCBIR

Page 8: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

What does MetricSPlat do?

It eases the process of raising CBIR metric spaces: It is modular (dlls) It uses a well-defined interface (arboretum) It provides a prototyping subproject for quick

development of new functionalitiesBesides, MetriSPlat is aided with

visualization techniques to:Monitor the datasetMonitor the result of similarity queries

Page 9: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Metric StructuresMetric Structures

and and

Distance Distance FunctionsFunctions

in MetricSPlatin MetricSPlat

Page 10: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Metrics Structures andDistance Functions

New distance functions and metric structures can be accomplished with a prototyping subproject

Page 11: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Metrics Structures andDistance Functions

New distance functions and metric structures can be accomplished with a prototyping subproject

L2=Euclidiana

r

L0=LInfinity=Chebychev

L1=Manhatan

Page 12: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Features vectors as parallel coordinates

Demo

Page 13: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

FeaturesFeatures

ExtractionExtraction

Page 14: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Features extraction

Currently: no support for features extraction

But there is support for extracted features!

Page 15: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Features extraction

Support for extracted featuresA simple way of loading data

Embedded database driver: OpenOffice, MSAccess

File: Features.mdb, placed at the application’s directory

Required fields:NAtributos: the number of attributes to considerCOD: uniquely identifies the records

Page 16: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Visualization SupportVisualization Support

Page 17: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

To doTo do

Page 18: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

To do

Modularize features extraction

Modularize media reproduction

Aid support for relevance feedback

Aid support for CBIR evaluation: Automatic Precision Recall Ranking function

Page 19: Metric s plat - a platform for quick development testing and visualization of content-based retrieval techniques

- XXIV Simpósio Brasileiro de Banco de Dados 2009 -

Thank youThank youhttp://www.icmc.usp.br/~junio

Link “MetricSPlat”

[email protected]


Recommended