I am studying how HTK Tools works with handwriting recognition. Following the ICFHR–2010 TUTORIAL I run examples for "Spanish-Numbers" corpus and received the resulting HMMs (files stored in folder hmm and listed in HMMsList), and res32.mlf with results of recognition received with HVite. Also I have master label file SamplesRef.mlf.
And now I want to see recognition results statistics, i.e. studying HResults tool.
When I run HResults as
HResults -I SamplesRef.mlf HMMsList res32.mlf
I see
====================== HTK Results Analysis =======================
Date: Tue Mar 31 15:21:11 2015
Ref : SamplesRef.mlf
Rec : res32.mlf
------------------------ Overall Results --------------------------
SENT: %Correct=0.00 [H=0, S=2, N=2]
WORD: %Corr=77.78, Acc=77.78 [H=7, D=0, S=2, I=0, N=9]
===================================================================
But if I add option -p in order to have confusion matrix I see the following error message:
~/icfhr$ HResults -p -I SamplesRef.mlf HMMsList res32.mlf
ERROR [+3331] Index: Label millones not in list[0 of 19]
FATAL ERROR - Terminating program HResults
I understand that message means that there is no HMM with name "millones" and I found that in my res32.mlf samples looks like:
"’*’/210341.rec"
mil
seiscientos
cincuenta
y
siete
millones
.
If I change res32.mlf with text editor to res33.mlf with content like:
"’*’/210341.rec"
m
i
l
s
e
i
s
c
i
... and so on.
And use samples.mlf (instead of SamplesRef.mlf) which inside looks like:
"*/210341.lab"
m
i
l
@
q
u
i
n
i
e
n
t
o
s
@
c
... and so on.
I have the desired result:
~/icfhr$ HResults -p -I samples.mlf HMMsList res33.mlf
====================== HTK Results Analysis =======================
Date: Tue Mar 31 15:35:42 2015
Ref : samples.mlf
Rec : res33.mlf
------------------------ Overall Results --------------------------
SENT: %Correct=0.00 [H=0, S=2, N=2]
WORD: %Corr=79.63, Acc=77.78 [H=43, D=5, S=6, I=1, N=54]
------------------------ Confusion Matrix -------------------------
a c d e i l m n o s t u v y Del [ %c / %e]
@ 0 0 0 0 0 1 1 0 0 0 0 0 0 0 5 [ 0.0/3.7]
a 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0
c 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0
d 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
e 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0
i 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0
l 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0
m 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0
n 0 1 0 0 0 0 0 6 0 0 0 0 0 0 0 [85.7/1.9]
o 0 0 0 0 0 0 0 0 4 0 0 0 0 0 0
q 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 [ 0.0/1.9]
s 0 0 0 0 0 0 0 0 0 4 0 0 0 0 0
t 0 0 0 0 0 0 0 0 0 0 4 0 0 0 0
u 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 [50.0/1.9]
v 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0
y 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 [50.0/1.9]
Ins 0 0 0 0 0 0 0 0 0 1 0 0 0 0
===================================================================
So, the main question is:
What is the simplest way (without text editor) to make mlf-files adapted for making confusion matrix?
(I suppose I miss some option of some HTK tool… but which tool and which option?)
Any useful ideas would be highly appreciated.
