| 12 | |
| 13 | * an approach: |
| 14 | |
| 15 | {{{ |
| 16 | Date : Wed, 6 Dec 2006 21:27:29 +0800 |
| 17 | De : "Xie Zhiqing" <kramxxx@gmail.com> |
| 18 | À : htk-users@eng.cam.ac.uk |
| 19 | Objet: [HTK-Users] Lightly supervised acoustic training |
| 20 | |
| 21 | Hi, my name is Mark and I am a student from Singapore. Currently I am |
| 22 | working on a project on speech recognition, specifically on trainning |
| 23 | portion of the system. From my knowledge, basically what I am |
| 24 | supposed to do is to train the system on a small amount of manually |
| 25 | transcribed speech (.wav and .lab) and then use it to transcribe a |
| 26 | larger amount of untranscribe speech (only with .wav). If the |
| 27 | confidence level is high enough, then it will be added into the |
| 28 | trainning data and the process will be run iteratively until all the |
| 29 | untranscribe speech is added to the trainning data. Is this method |
| 30 | correct? |
| 31 | |
| 32 | >From what I gather from the HTK book, using Hvite it will output |
| 33 | transcriptions for the raw speech. The format is as such => start |
| 34 | time , end time , phoneme and the total log probability. Is there a |
| 35 | connection between the confidence measure and the total log |
| 36 | probability? I am currently using Matlab to implement HTK. |
| 37 | |
| 38 | Are there any good sites that can explain the process in more layman |
| 39 | terms? |
| 40 | |
| 41 | }}} |