topic model 0.002630323
same model 0.0022961559999999997
image feature 0.002290763
annotation model 0.002200909
image features 0.002122975
latent model 0.002118039
lda model 0.002113634
other image 0.002091351
image annotation 0.002083939
image caption 0.0020663820000000003
relevance model 0.002030045
model description 0.002014876
our model 0.00197222
model our 0.00197222
model parameters 0.00196615
image classification 0.001951676
document word 0.0019439869999999999
model top 0.0019426539999999998
image descriptions 0.001934993
image regions 0.001929673
multinomial model 0.001926498
test image 0.001914146
relevance image 0.0019130750000000002
automatic image 0.001896216
annotated image 0.00189131
model cap 0.001888891
extended model 0.0018886279999999998
model access 0.001886975
evance model 0.0018833909999999999
model benefits 0.0018833909999999999
current image 0.001841041
image retrieval 0.001840311
image segmentation 0.001840193
robust image 0.001811368
image anno 0.001799578
traditional image 0.001792752
left image 0.00178275
image annota 0.001782429
accurate image 0.0017799830000000002
image collections 0.0017727790000000002
unknown image 0.001766902
center image 0.001766182
model 0.00167732
document words 0.0016597180000000001
word sense 0.0015806359999999998
same topic 0.001571839
image 0.00156035
word list 0.001505826
word prob 0.001500261
annotation words 0.0014848399999999999
caption words 0.001467283
word frequen 0.001451971
content words 0.001411299
tion words 0.001335799
training data 0.001321685
test words 0.001315047
annotation models 0.001312403
feature vector 0.0013091919999999998
annotated words 0.001292211
test data 0.001277584
strong topic 0.0012757369999999999
output words 0.001259465
visual features 0.001257168
news images 0.001241111
data our 0.001218688
data sets 0.0011947820000000001
likely topic 0.001190069
title words 0.001182689
same features 0.001181461
noisy data 0.001180639
broad topic 0.001171879
topic categories 0.001171218
topic indicators 0.001170897
notation words 0.001169819
topic mix 0.001160629
news dataset 0.001155534
data acquisition 0.0011475230000000001
variable models 0.001147238
relevance models 0.001141539
news documents 0.0011378739999999999
annotation task 0.001126094
average document 0.001091339
feature vectors 0.001053517
news database 0.001050445
graphical models 0.001050275
additional information 0.001024248
good feature 0.0010216169999999998
mixture models 0.001021567
news source 0.001017552
important information 0.001007577
feature selection 0.001004876
news articles 0.001001438
information retrieval 9.98925E-4
tation models 9.98655E-4
notation models 9.97382E-4
news website 9.938339999999999E-4
news article 9.91828E-4
document collection 9.81985E-4
news arti 9.628659999999999E-4
bbc news 9.626559999999999E-4
