Semi-automatic subtitle editor (didactical project for Cognitive Services course)
The aim of this project is to implement a smart subtitle editor for english-language videos. Writing subtitles from scratch can be a very boring activity, because the writer has to split all speeches into sentences and, for each of these, he has to identify the time position inside the video and the text said. Instead with this project (from now on called Subeeper), all these operations are done by the program automatically and the user has only to check and edit the text suggested for each sentence, which is (as well as the text) identified and time located by using audio captioning and cognitive services APIs.
The program interface need to be very simple to use. It must give the possibility to view the video under processing, to have information about the current time position and to edit the text which will be shown as subtitle. A mockup is reported here:
The architecture follows MVC design pattern. A simple class diagram is reported to show the main structure
- ffmpy and FFmpeg (https://www.ffmpeg.org/)
- PyQt5 (http://pyqt.sourceforge.net/Docs/PyQt5/installation.html)
- VLC media player (https://www.videolan.org/vlc/index.it.html)