LIRE图像检索引擎Lire - Lucene Image REtrieval

Lire - Lucene Image REtrieval

The LIRE (Lucene Image REtrieval) library a simple way to create a Lucene index of image features for content based image retrieval (CBIR). There is no complete list of features, but these are some of them:

• ScalableColor, ColorLayout and EdgeHistogram MPEG-7
• CEDD and FCTH (contributed by Savvas Chatzichristofis)
• Color histograms (HSV and RGB), Tamura & Gabor, auto color correlogram, JPEG coefficient histogram (common global descriptors)
• Visual words based on SIFT and SURF
• Visual words based on SIMPLE
• Approximate fast search based on hashing and metric indexing.

Furthermore, methods for searching the index based on Lucene are provided.

The LIRE library started out as part of the Caliph & Emir project and aimed to provide the CBIR features of Caliph & Emir to other Java projects in an easy and light weight way. In the meantime it has turned out as big and interesting project itself.

With Lire you can easily create an index and search through the index. LIRE 1.0 also supports local features based on bag of visual words and the SIMPLE approach, see Builders.

• How to create an index with Lire?
• How to search through the index with Lire?
• Frequently Asked Questions

I recommend to start with taking a look at the SimpleApplication package of LIRE, which covers the most needed stuff including indexing, search and extraction of image features for use in other applications. It's also a good idea to work on the current SVN version of LIRE: How to check out and set up LIRE in the IDEA IDE.

Note at this point that LIRE comes with Apache Ant build files, named build.xml. You can use the tasks to create the jar from the source code as soon as you have Ant installed, or you are using an IDE prepared for that, like IDEA, Eclipse or NetBeans. Apache Ant can be found at the Apache Ant Project Page

If you are searching for the Solr plugin of LIRE ... it's still under construction. Some global features are working fine and its based on Solr 4.10.2. It can be found at BitBucket. It has been reported working on distributed installations.

How does Lire actually work?

Lire employs global image features for content based image retrieval. For more information on the underlying methods and techniques you should consult the basic literature on content based images retrieval:

• Visual Information Retrieval using Java and LIRE (Lux & Marques, 2013)
• Image Retrieval: Ideas, Influences, and Trends of the New Age (Datta et al., 2008)
• Content-Based Image Retrieval at the End of the Early Years (Smeulders et al., 2000)

Further it uses the Java search engine Lucene to provide

• linear search (opening each and every indexed document and comparing it to the query feature)
• approximate indexing based on metric spaces based on the work of G. Amato
• approximate indexing based on locality sensitive hashing

Performance

Parallel indexing with the ParallelIndexer running with 8 threads on a AMD A10 with 4 cores and 4.4 GHz, Windows 7 64 bits extracting 7 features at once including hashing is down to ~180 ms per image. On a Intel Core i7, ie. the 4770K, it runs a lot faster, using an SSD then speeds up the process even more. Extracting single features with the ParallelIndexer is on a core i7 typically faster than 1 MP images can be read from a (magnetic) hard disk.

Search is a matter of index size and is down to a few ms for 100k and less images, and increases linearly.

----------------------------------------------------------------------------------------------------------------

Lire

An Open Source Java Content Based Image Retrieval Library

The LIRE (Lucene Image REtrieval) library provides a simple way to retrieve images and photosbased on their color and texture characteristics. LIRE creates a Lucene index of image features for content based image retrieval (CBIR). Several different low level features are available, such as MPEG-7 ScalableColor, ColorLayout, and EdgeHistogram, Auto Color Correlogram, PHOG, CEDD, JCD, FCTH, and many more. Furthermore simple and extended methods for searching the index and result browsing are provided by LIRE. LIRE scales well up to millions of images with hash based approximate indexing. The LIRE library and the LIRE Demo application as well as all the source are available under the Gnu GPL license.

Documentation and help are available

• on the developers wiki
• on the developers mailing list
• on the Google Code page of LIRE
• in the LIRE book on Amazon or Morgan & Claypool

Please drop me a line on the mailing list if you use LIRE in your project. I would appreciate feedback very much and I will link reference projects.

If you like this work and you want to support the development of Lire, consider to donate to the project.

Consulting & Services

You plan to work with LIRE and you are stuck? I offer additional consulting services for LIRE, visual information retrieval and multimedia management. Note that the software is free, but I can help you with questions like:

• Which ones are the right features for me?
• How do I implement a service like XXX?
• What is it LIRE can / cannot do for me?
• How to integrate more intelligent processing methods like LSA or PCA?
• What are the technical implications and limits when using LIRE?
• … and many more.

Please contact me for further information on consulting at . All non consulting-related requests should be directed to the mailing list.

About the citation of Lire

We kindly ask you to refer the following paper in any publication mentioning Lire:

Lux Mathias, Savvas A. Chatzichristofis. Lire: Lucene Image Retrieval – An Extensible Java CBIR Library. In proceedings of the 16th ACM International Conference on Multimedia, pp. 1085-1088, Vancouver, Canada, 2008

• Download paper here (or visit the ACM page)
• Download BibTeX

… or of course the book Visual Information Retrieval with Java and LIRE.

Downloads

• Lire downloads
• LireDemo (GUI application to try out)
• Lire nightly build – a fresh build each day from the SVN

Lire Demo

Included with the release the LIRE Demo found its way into the LIRE package. It provides a simple GUI interface for

• Indexing photos
• Searching photos
• Browsing the created index
• Creating mosaic images based on indexed images (thanks to Lukas Esterle & Manuel Warum for contributed the mosaic-ing code.)

The mosaic creation is described in detail in this post. You can also find a flash based demo of the mosaic creation abilities of Lire.
Visit sourceforge.net to download LireDemo. Note that Java 1.6 is needed for using LIRE Demo.Installation instructions:

• Download LIRE Demo
• Unzip (or untar) the downloaded files.
• Ensure that you haveJava 1.6 installed.
• Start the demo with “java -jar LireDemo.jar” (or with a double click if you are using Windows)

Online LIRE Demo

An online demo of LIRE is available at http://demo-itec.uni-klu.ac.at/liredemo/ featuring approximate search in 300,000 images.

from: http://www.semanticmetadata.net/wiki/

http://www.semanticmetadata.net/lire/