[Gzz-commits] manuscripts/Paper paper.tex

[Janne V. Kujala]

CVSROOT:        /cvsroot/gzz
Module name:    manuscripts
Changes by:     Janne V. Kujala <address@hidden>        03/03/17 11:51:21

Modified files:
        Paper          : paper.tex 

Log message:
        evaluation

CVSWeb URLs:
http://savannah.gnu.org/cgi-bin/viewcvs/gzz/manuscripts/Paper/paper.tex.diff?tr1=1.25&tr2=1.26&r1=text&r2=text

Patches:
Index: manuscripts/Paper/paper.tex
diff -u manuscripts/Paper/paper.tex:1.25 manuscripts/Paper/paper.tex:1.26
--- manuscripts/Paper/paper.tex:1.25    Mon Mar 17 08:08:25 2003
+++ manuscripts/Paper/paper.tex Mon Mar 17 11:51:21 2003
@@ -1077,32 +1077,46 @@
 
 \subsection{Recognizability and memorizability}
 
-JVK
-
-- lots of texture perception work on texture discrimination 
 
-- we need also memorization; therefore, careful evaluation is needed 
-
-- our textures are on a higher level, more like complete pictures
+JVK
 
-- experiments on black-and-white (faces,) ink blots, and snow crystals
-  \cite{goldstein71visualrecognition} show that
-  complex pictures can be remembered and recognized and that recognition
-  performance decreases very little over time.
-
-- pilot test: 15 textures or pure colors shown sequentially, 5 seconds each.
-  recognition tested by showing the 15 seen textures mixed with 15 new
-  textures in a random order, subject answers old/new. 
-
-- results: for textures, 80 percent correct (12/15 correctly recognized
-  as seen and 12/15 correctly identified as not seen).
-  for pure colors, 53 percent correct (13/15 correctly recognized
-  as seen and 3/15 correctly identified as not seen).
-
-- the experiment only measured the ability to dicriminate previosly seen
-  textures from unseen textures. However, our experience shows that
-  at least the most recurring textures can easily be associated to the 
-  document content.
+There has been
+lot of texture perception work on texture discrimination.
+However, in our application texture discrimination is not as
+much of an issue as memorizability and recognizability of
+previously seen textures.
+Furthermore, our textures are on a higher level,
+more like complete pictures than the usually studied microstructure.
+
+Experiments on black-and-white %(faces,) 
+ink blots, and snow crystals
+\cite{goldstein71visualrecognition} show that
+complex pictures can be remembered and recognized and that recognition
+performance decreases very little over time.
+
+We have conducted a pilot experiment (with one subject) 
+in a similar setting comparing
+the recognition performance of our textures with pure color backgrounds.
+First, 15 textures (or pure colors) were shown sequentially, 5 seconds each.
+Then, recognition was tested by showing the 15 seen textures mixed 
+with 15 unseen textures in a random order, and the subject answered 
+old or new. 
+The results were as follows: 
+for textures, 80 percent correct (12/15 correctly recognized
+as seen and 12/15 correctly identified as not seen);
+for pure colors, 53 percent correct (13/15 correctly recognized
+as seen and 3/15 correctly identified as not seen).
+
+Of course, this result is not statistically relevant,
+but in conjunction with our experiences, it suggests that the
+textures are quite recongizable and that they are clearly
+superior to pure colors in terms of recognizability.
+
+Furthermore, the experiment only measured the ability to 
+recognize previosly seen textures within unseen textures. 
+Our experience shows that in practice,
+at least the most recurring textures are not only recognized but can 
+also easily be associated to the document content.
 
 - as per Zipf's law [XXXref], it suffices to learn the textures 
   of the most recurring documents.