CBAS is an image search engine for finding art. It emerged from some work done on Content Based Image Recognition and experiments with realtime php image recognition.
Unlike traditional search engines the input for a CBAS search is actually an image rather than text. This is not useful if you are searching for a particular work by a known artist. A mainstream search engine would be best suited to that kind of search. Where CBAS is useful is if you have found an image but you know nothing about it. That image can be used as the source and hopefuly if a match is discovered in the database, it will lead you to information about the picture.
These essays and experiments were written for the seeker (~s~) community at Searchlores, a bottomless pit of web knowledge. If you are interested in techniques for searching the internet, prepare to be amazed and overwhelmed by the power these pages will give you.
It is for this reason that CBAS is dedicated to Searchlores, Fravia and the community there.
CBAS is still very much under development. The code and index is constantly under revision and as such predictable results cannot be guaranteed.
Currently the searchable index is limited to images found at the Web Gallery of Art. The gallery is generally limited to European art produced within the 12th to 18th centuries. It is a valuable resource and worth exploring in it's own right. This project is not affiliated with WGA in any way.
There are limitations to the current implementation. As the comparison is based on colour frequency, images which have had colours filtered may not match. Another significant limitation is if an image has been cropped. The regions will not match up and most likely colours in the source and target regions will not match.
Ongoing development hopes to address these issues where possible.
CBAS is written entirely in PHP and the code will be(after some polishing and tidying) released as free software licensed under the GNU GPL.
Search - Help - Contact - Example