[Sigcr-l] re: similar studies

[beebe at indiana.edu]

Hello Corrine,
I have been out-of-the loop for the past year, but have managed to plug along 
at my picture-description transcriptions. Hope to finish by Christmas and do 
the analysis and write-up by next summer. Luckily someone passed on this 
current sigcr-l posting to me, I didn't know about the list!

By image structure I mean the visual components of the picture (without 
interpretation as higher-level concepts) e.g., color, shape, direction 
relations, pattern, etc. I specifically chose images that were not readily 
identifiable as specific objects: celullar structures, aerial photos of 
unrecognizable landscapes, abstract art with no identifiable objects. Forty 
subjects worked in pairs , one describing the image for the other to draw. The 
goal of the exercise was to create a drawing that could be used to find that 
particular picture out of a collection of hundreds. The drawer could not see 
the picture until the describe-draw task was completed.

Though I will stick to the vocabulary aspect first, I am also excited about the 
descriptive process and the implications it might have for both CBIR 
interfaces, and image description for the blind when tagging HTML.

Thanks for the reference to last years ASIS paper, I'll look it up. Ordered 
your book today, can't wait to read it! 

Best, Caroline