I am working on a plugin that connects to a speech recognizer. Is there any interest in this in the Nutch community?
It would not be a speech recognizer plugin as such though anyone interested in doing this should probably look at the open-source Sphinx4 project for that purpose (personally I don't think it's a good idea to spawn a new speech recogniser each type there is some audio to be parsed so it would be best run as a server). Rather, what I'm thinking of is a plugin that connects to a web-service: your Nutch instance sends the audio (from mp3/wav etc. MIME types) to be recognized and gets back a text stream that can be indexed by Nutch. The web-service would probably be free for small volumes and fee-based for larger volumes or high throughput since it's pretty compute expensive running decent quality speech recognisers. How much interest is there in indexing audio/video in this manner? -Ed On 3/21/07, karl wettin <[EMAIL PROTECTED]> wrote:
21 mar 2007 kl. 11.27 skrev Anton Potekhin: > Does nutch have a video search solution? That I doubt. > Does anybody do anything about this problem? > Can I read about it anywhere? It really depends on what you want to use as a query and what in the video you are looking for. Pattern recognition in images is fairly well documented. http://citeseer.ist.psu.edu/cis?q=video+search&cs=1 -- karl
