|
|
||
|
You are browsing the image database which currently contains 754033 images, provided by Matton AB in Stockholm.
The ImBrowse search engines have been developped by Thanh H. Bui and Reiner Lenz
within the "Content Based Search in Image and Video Databases" project, funded by Vinnova.
An earlier version of this search engine (written by Dr. Linh V. Tran in March 2002 as part of
his doctoral thesis) can be found at the site of Linköping
University Electronic Press. (This project ended in 2005)
If you want to skip this brief tutorial, just click the following button to directly go to the ImBrowse. Welcome! |
||
|
This is a fullscreen snapshot of ImBrowse CBIR system. ImBrowse currently provides the following modes to browse the database:
|
|
![]()
|
You can search by entering a keyword in the textbox or you can click on a keyword in the keyword list. Entering a valid URL for an image on the web will use this image as query. You can also use the bookmarklet for collecting images from a web page, and then use one of these as URL image query. Keyword search has higher priority than URL-search. Pressing the Info button displays information related to the current search query. You can also choose how many images are going to be displayed on one page.
Clicking the Random link will retrieve a new random selection of images. If the database contains keywords for the current query image then these keywords will be displayed. Clicking any keyword will launch a keyword search query. |
|
|
Currently we provide five different content-based search engines. You can select the search engine for the next query by pressing one of these radio buttons. A search can be based on color and/or texture properties of the image. The different search modes can also be selected by pressing one of the numbered buttons below each image on the page. In this case the selected image becomes the query image and the database is searched by the corresponding search engine. You can also customize the search by changing the slider control to give more weight on the favourable features. The slider is located between: (T) stands for texture feature; and (C) stands for color feature.
|
|
|
Bellow each image you find sevaral buttons, the © button will connect you to the original image on the Matton site.
The numbered buttons will launch a content-based search for the image with different search modes. The number refers to the search modes listed in the upper right corner. The "K" button will retrieve keywords for the selected image.
|
|
|
You can click the Info button to display the current image and the graph of the matching scores for the first 100 matched images.
|
|
|
Technical responsible: Martin Solli
Digital Media, ITN, Linköping University Last updated: 2006-09-10 |
|
|