Hello, I am new to the voice recognition domain and I am using Sphinx3. I want to achieve recognition accuracy at about 95 % with small interactions as follows.
Sentences in Corpus:
Shepherd03, this is Fox01, over
Fox01, this is Shepherd03, over
Shepherd03, this is Fox01, SITREP, mission # (N), current location, capping Caliente High as fragged, over
Shepherd03: Fox01, this is Shepherd03, acknowledge SITREP, Fox01 capping at Caliente High as fragged, mission # (N), over
I run this corpus thru the LMTool, then thru lm3g2dmp tool to get my DMP file. How can I achieve this?
Any help will be appreciated.
Lawrence
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
For an up coming demo I have a small interaction that I would like to recognize but for the future I have lots of interactions to account for. What would be my best long term solution, keeping in mind that we would like to have 95 % and above recognition accuracy?
Lawrence.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
if you just want to recognize below than 50 words, you may want to train you own model for a better accuracy. if you only want the engine to recognize just this 4 sentence, you may want to used a grammar instead of n-gram language model.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello, I am new to the voice recognition domain and I am using Sphinx3. I want to achieve recognition accuracy at about 95 % with small interactions as follows.
Sentences in Corpus:
I run this corpus thru the LMTool, then thru lm3g2dmp tool to get my DMP file. How can I achieve this?
Any help will be appreciated.
Lawrence
Thank you for your response!!
For an up coming demo I have a small interaction that I would like to recognize but for the future I have lots of interactions to account for. What would be my best long term solution, keeping in mind that we would like to have 95 % and above recognition accuracy?
Lawrence.
if you just want to recognize below than 50 words, you may want to train you own model for a better accuracy. if you only want the engine to recognize just this 4 sentence, you may want to used a grammar instead of n-gram language model.