https://github.com/luser/auto-caption
Produce captions for videos using PocketSphinx speech recognition
https://github.com/luser/auto-caption
Last synced: about 1 year ago
JSON representation
Produce captions for videos using PocketSphinx speech recognition
- Host: GitHub
- URL: https://github.com/luser/auto-caption
- Owner: luser
- License: other
- Created: 2015-11-18T02:21:57.000Z (over 10 years ago)
- Default Branch: master
- Last Pushed: 2016-01-15T21:33:59.000Z (over 10 years ago)
- Last Synced: 2025-03-08T00:04:02.203Z (over 1 year ago)
- Language: Python
- Size: 7.81 KB
- Stars: 8
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
This project contains two scripts that use [PocketSphinx](https://cmusphinx.sourceforge.net/) to produce captions
from a video file using speech recognition. The dependencies are a bit tricky,
a Dockerfile is provided to produce a working environment. Specifically,
the script currently relies on [an unlanded patch](https://github.com/luser/pocketsphinx/commit/18f6755caafd04726e76569aa7daa7c6211ea05e) to the PocketSphinx
Gstreamer plugin.
`caption.py` takes a media file and generates a caption file. You can test
this script with the pre-built docker image `luser/auto-caption:0.2`, for example:
```
docker run -t luser/auto-caption:0.1 ./run.sh https://people.mozilla.org/~tmielczarek/test-long.wav
```
Will produce captions on stdout.
`adapt-from-captions.py` takes a media file, a manually corrected captions
file, and a PocketSphinx acoustic model, and adapts the model by feeding it
the matched input audio and corrected text. It will output
`updated-model.tar.gz` in the working directory if it succeeds.
Any copyright is dedicated to the Public Domain.
http://creativecommons.org/publicdomain/zero/1.0/