Automatic Speech Recognition Software

I am looking for a specialist in Automatic Speech Recognition to build some speech recognition software for me. I want the software to be able to analyse a sound file, and produce a transcription of the words in a text file, in the first instance. Ultimately, I would like the data to be stored in a database and an API be available to integrate the ASR software with other software.

The analysis of the audio file should be quick - I appreciate that this is not a simple thing, but the faster the file can be processed the better. The audio files will be of different people speaking in a conversational tone, and so the software will need to cope with that. Accuracy of at least 80% is expected, higher if possible. At this time, transcription in English is all that is required. The output will need to be time stamped so that, for example, it shows that at 1 minute 25 seconds, the word "computer" was spoken. The software should be able to recognise that a new file has been received, load it, process it and output the results automatically. Any errors or issues need to be reported (perhaps via email) to the admin.

I have suggested the language to be used is python, mainly because I like python, and it is powerful and flexible. If you have a convincing argument as to why we should use something else, I am open to being persuaded. I am only looking to hire someone who has experience in this field, as it is complex, and so I would like evidence of other similar projects that have been successfully completed.

The software will need to be installed on one of my servers (running Ubuntu [url removed, login to view] at the moment), access to which can be provided remotely. The successful applicant will be required to sign an NDA.

Python Transcrição

ID do Projeto: #5977904

Sobre o projeto