img(Anaktisi) Image Search Engine

February 27, 2008 on 10:00 am | In Imaging, Multimedia, Retrieval, search |

img-search.jpgSince Blobworld was down I was hoping for a new online CBIR system to show in lectures. Last week I received word about the img(Anaktisi) image search engine from Savvas Chatzichristofis. This search engine is based on 2 new content based descriptors that combine color and edge features. The results look promising!

There is also an offline sketch based retrieval system being developed. A screencast can be found on YouTube.

Related Links



5 Comments »

RSS feed for comments on this post. TrackBack URI

  1. Which systems (etither online or offline) are you currently demonstrating in your lessons?

    Comment by Jose Maria Gomez — February 27, 2008 #

  2. The online one for sure, I’ve to try the offline version first :) (haven’t done this yet)

    Comment by Mathias Lux — February 28, 2008 #

  3. I am afraid I was not clear. I meant: which other systems, apart form img(Anaktisi), are you using in the lessons?

    I wanted to suggest several, but it is very likely you know all of them, and I do not like to seem stupid (as I have found all of them the same place I found LIRE, that is, the Content Based Image Retrieval article at the Wikipedia). Anyway, I would like to suggest:

    Flikr Suggestions, the server-side demonstration of imgSeek on Flikr.
    Retrievr, that roughly implements the same Van Harr wavelets algorithms as imgSeek.
    ALIPR, much more sophisticate than the previous ones, and that aims at concept-tagging of images (not so content-based retrieval, but much powerful regarding image processing).

    Hope not to be redundant :-)

    Comment by Jose Maria Gomez — February 28, 2008 #

  4. Retrievr is one of my favourites because of the simplicity of the approach. The flash interface is super easy to use. ALIPR however was new to me. I like the results but it is not clear to me if this is an approach combining user input and content based descriptors. In this case the results are not as surprising as they seem :)

    Comment by Mathias Lux — February 28, 2008 #

  5. As far as I understand what I have read about ALIPR, the retrieval is by the vocabulary, but the system makes analysis of mistakes to learn the properties of the concepts. So not properly CBIR :-)

    Comment by Jose Maria Gomez — February 28, 2008 #

Leave a comment

XHTML: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

© 2004-2007 by Mathias Lux
>> Contents of this page are licensed under the CreativeCommons Attribution 2.5 license <<