1242 |
Experiments on black-and-white %(faces,) |
Experiments on black-and-white %(faces,) |
1243 |
ink blots and snow crystals |
ink blots and snow crystals |
1244 |
\cite{goldstein71visualrecognition} show that |
\cite{goldstein71visualrecognition} show that |
1245 |
complex pictures can be remembered and recognized and that recognition |
complex, unfamiliar pictures can be remembered and recognized |
1246 |
|
and that recognition |
1247 |
performance decreases very little over time. |
performance decreases very little over time. |
1248 |
|
|
1249 |
We have conducted a pilot experiment (with one subject) |
We have conducted an experiment in a similar setting comparing |
1250 |
in a similar setting comparing |
the recognition performance of our textures with solid color backgrounds. |
1251 |
the recognition performance of our textures with solid color backgrounds |
Five participants performed the experiment in both conditions |
1252 |
(using colors from the same distribution that is used for the textures). |
(in random order). |
1253 |
First, 15 textures were shown sequentially, 5 seconds each. |
|
1254 |
Then, recognition was tested by showing the 15 seen textures |
15 target backgrounds and 15 distractor backgrounds were randomly |
1255 |
with 15 unseen textures in a random order, and the subject answered |
chosen for both conditions. |
1256 |
old or new. The test was repeated with solid color backgrounds |
The distribution of the solid colors was the same that is used for |
1257 |
for comparison. |
the colors of textures with the difference that the highest |
1258 |
|
lighness tail was de-emphasized to increase the otherwise low |
1259 |
The results were as follows: |
discriminability of very light, unsaturated colors. |
1260 |
for textures, .3 bits per trial of the old/new information got through |
|
1261 |
(80 percent correct: 12/15 correctly recognized as seen and |
First, the 15 target backgrounds were shown sequentially, |
1262 |
12/15 correctly identified as not seen); |
in a random order, 5 seconds each. |
1263 |
for solid colors, only .003 bits per trial got through |
Then, recognition was tested by showing the 15 target backgrounds |
1264 |
(53 percent correct: 13/15 correctly recognized as seen and |
and 15 distractor backgrounds in a random order and |
1265 |
3/15 correctly identified as not seen). XXX |
having the participant answer ``old'' or ``new'' for each one. |
1266 |
|
The time for answers was not limited. |
1267 |
|
|
1268 |
|
The results are summarized below |
1269 |
|
(percentage of trials and reaction time in seconds): |
1270 |
|
\begin{center} |
1271 |
|
\begin{tabular}{c|cc|cc} |
1272 |
|
& \multicolumn{2}{c|}{textures \% (RT)} & \multicolumn{2}{c}{colors \% (RT)} \\ |
1273 |
|
& correct & incorrect & correct & incorrect \\ |
1274 |
|
\hline |
1275 |
|
old & 84 (3.0) & 16 (5.6) & 43 (3.4) & 57 (4.0) \\ |
1276 |
|
new & 72 (2.6) & 28 (5.5) & 71 (2.8) & 29 (4.1) \\ |
1277 |
|
\hline |
1278 |
|
total & 78 (2.8) & 22 (5.5) & 57 (3.1) & 43 (4.1) |
1279 |
|
\end{tabular} |
1280 |
|
\end{center} |
1281 |
|
|
1282 |
%80, 53 fail to bring out the results correctly. Should |
%80, 53 fail to bring out the results correctly. Should |
1283 |
%use mutual information or something like that. In colors, the subject |
%use mutual information or something like that. In colors, the subject |
1284 |
%was basically guessing!!! |
%was basically guessing!!! |
1285 |
|
|
1286 |
Thus, the subject was able to recognize textures ($p < .001$) |
% RT less for correct responses: [F(1,4) = 5.2, p = .082] |
1287 |
while basically guessing the solid colors. |
Thus, the recognition performance of the textures was good while |
1288 |
Of course, the quantitative results have little statistical significance, |
the solid colors do not have enough variation for unambiguous |
1289 |
but in conjunction with our experiences, they do suggest that the |
recognition (without additional cues). |
1290 |
textures are quite recongizable while solid colors do not have |
The observed difference in recognition performance is statistically |
1291 |
enough variation for unambiguous recognition (without additional cues). |
significant [F(1,4) = 19.0, p = .012]. |
1292 |
|
|
1293 |
The experiment measured the recognition of only 15 textures, |
Of course, the experiment measured the recognition of only 15 textures, |
1294 |
and surely it would be harder to remember, e.g., 100 textures. |
and surely it would be harder to remember, e.g., 100 textures. |
1295 |
However, the user does not have to remember all the textures; |
However, the user does not have to remember all the textures; |
1296 |
it suffices to learn the textures of the most often used documents. |
it suffices to learn the textures of the most often used documents, |
1297 |
|
see Fig.~\ref{fig-zipf}. |
1298 |
Many studies of web cache statistics |
Many studies of web cache statistics |
1299 |
have shown that file popularity approximately follows Zipf's law |
have shown that file popularity approximately follows Zipf's law |
1300 |
(see, e.g. [xxxref]) so that a small number of documents accounts |
(see, e.g. [xxxref]) so that a small number of documents accounts |