One of my classmates in Prof. Macek’s new media course requested I post a link to my final paper on my blog, so here you go: “Captioning & accessible video on the Web”.
I was also asked about whether or not any progress had been made towards a useful, automatic captioning system based on speech recognition. I’m too sure about developments in this area, but I did find a short paper by Trmal, et al. describing an automatic speech recognition–based system for captioning parliamentary meetings in the Czech Republic. The system appears to be reasonably successful; however, the paper notes that captioning parliamentary sessions is, in some ways, unusually easy due to the carefully enforced rules of procedure which ensure that a speaker’s voice is generally unimpeded by the voices of others and by background noise.
Pingback: “I’ve got pi here, and I’ve got pi there; I’m pi-winning!” | Jonathan Rascher