The Germans have created a library of drunken audio recordings.
Epigraph:
- Drank?
- Did not drink!
- Say Gibraltar.
- Drank.
The topic is a bit curious, but it seems to me that it’s these lately that are not enough for Habré. Especially on Friday.
')
So, German scientists from two universities in Munich created a database of audio recordings, which recorded the speech of 162 people intoxicated. The data was collected from 2007 to 2009 and now the language corpus is being created on the basis of the database ( an article about the buildings on the wiki) Alcohol Language Corpus (ALC).
The body is supposed to be used for machine learning so that automation systems (for example, in a car) automatically recognize the state of the host and block the control. Thus, it is supposed to fight unscrupulous human factor.
Prior to this, several car concerns worked on detecting drunken behavior. But then it was either about sensors detecting alcohol vapors (and sometimes falsely triggered in the presence of a drunken passenger), analysis of sweat on the speed shift knob, or cameras that gazed intently at the driver's face. But, apparently, no one has thought of speech analysis yet.
The cost of a set of records - about $ 1200 (an impressive amount for 162 structured speeches of German alcoholics :)
By the way, here is an example from the database.
The news is finished, you can put a minus for brevity.
- Drank?
- Did not drink!
- Say Gibraltar.
- Drank.
The topic is a bit curious, but it seems to me that it’s these lately that are not enough for Habré. Especially on Friday.
')
So, German scientists from two universities in Munich created a database of audio recordings, which recorded the speech of 162 people intoxicated. The data was collected from 2007 to 2009 and now the language corpus is being created on the basis of the database ( an article about the buildings on the wiki) Alcohol Language Corpus (ALC).
The body is supposed to be used for machine learning so that automation systems (for example, in a car) automatically recognize the state of the host and block the control. Thus, it is supposed to fight unscrupulous human factor.
Prior to this, several car concerns worked on detecting drunken behavior. But then it was either about sensors detecting alcohol vapors (and sometimes falsely triggered in the presence of a drunken passenger), analysis of sweat on the speed shift knob, or cameras that gazed intently at the driver's face. But, apparently, no one has thought of speech analysis yet.
The cost of a set of records - about $ 1200 (an impressive amount for 162 structured speeches of German alcoholics :)
By the way, here is an example from the database.
The news is finished, you can put a minus for brevity.
Source: https://habr.com/ru/post/245125/
All Articles