error model 0.00356215
word error 0.00343698
language model 0.00332772
translation model 0.002995109
ocr data 0.00275093
character error 0.0027151
level model 0.0026737139999999998
ocr error 0.00263584
previous model 0.002616733
source model 0.002584709
word level 0.002548544
channel model 0.00250611
original model 0.002462592
word sequence 0.002462386
length model 0.002457337
correct word 0.002451889
tion model 0.002444145
segmentation model 0.002424343
model component 0.002408906
guage model 0.002405507
model combination 0.002395954
nel model 0.002393267
mentation model 0.002392737
model fst 0.002392737
model fsa 0.002392737
word split 0.00235466
split word 0.00235466
word usage 0.002308667
conventional word 0.002275453
character errors 0.002241136
error correction 0.002175519
ocr errors 0.002161876
model 0.0021562
training data 0.0021161599999999997
ocr system 0.002098448
test data 0.002051468
input data 0.002013406
ocr correction 0.001999459
data wer 0.001925741
arabic data 0.001924106
error models 0.0019187050000000001
ing data 0.001909313
error reduction 0.001830382
character level 0.001826664
bible data 0.001757862
ocr process 0.001750324
first ocr 0.001748303
character sequence 0.001740506
source character 0.001737659
error rate 0.001731114
correct character 0.001730009
split error 0.00172958
relative error 0.001725057
translation system 0.0017074669999999998
ocr performance 0.001705643
automatic error 0.001690122
language models 0.0016842749999999998
error limit 0.001683325
error mod 0.001680691
ocr systems 0.001669986
english ocr 0.001669345
error rates 0.0016677039999999999
error reductions 0.001649399
character recognition 0.001647733
inverted error 0.001642239
character sequences 0.001627614
space character 0.001615064
ocr output 0.001602886
character shape 0.001600266
same language 0.001591808
character transformation 0.001579708
character distributions 0.001567305
optical character 0.001553966
ocr sequences 0.001548354
commercial ocr 0.0015479769999999999
inal character 0.001547391
character recognizers 0.001546438
observed character 0.001546438
correction correction 0.001539138
original ocr 0.001536282
new ocr 0.001534564
real ocr 0.001527233
bic ocr 0.001507547
ocr sys 0.001501862
ocr development 0.001501595
native ocr 0.001501555
language processing 0.0015008209999999998
ocr degradation 0.001496111
rewrite ocr 0.001486399
ocr post 0.001486031
systran ocr 0.001483079
natural language 0.0014801889999999998
new language 0.001476194
trainable ocr 0.001473059
trained ocr 0.001468574
ocr capabilities 0.001468448
density language 0.001458367
language dry 0.0014361159999999999
ral language 0.0014085579999999999
african language 0.001408012
