Live Speech Recognition
Speech recognition translates spoken information into digital text in real time
Fraunhofer
Developed by
License
Other
Intellectual property of Fraunhofer IAIS (closed source)
Main Characteristic
Live speech recognition reliably translates spoken information into digital text in real time.
Main characteristics:
- highly reliable speech recognition
- robust against noise, e.g. in an industrial setting
- can be combined with automatic speaker recognition
- language models available for German and English
- word and phoneme output to subsequent systems
Research areas
Physical AI
Technical Categories
Audio processing
Last updated
21.04.2023 - 14:18
Detailed Description
The provided container is packaged in the following asset deployed in the AI4EU Experiments Platform: Speech Recognition in AI4EU Marketplace
Trustworthy AI
The ASR service is (1) lawful, as it respects all applicable laws and regulations (e. g. software licenses of used open source components), especially it is GDPR-compliant, (2) ethical, as it pursues the ethical goal of making information from media data easily accessible in digital form to the datas' owner, (3) robust, from a technical perspective, especially as it is deployed in a "ready-to-use" Docker container, to make processing documents as simple as possible.
GDPR Requirements
The ASR service allows the user to translate spoken words from an audio stream into digital text. The software itself is GDPR compliant. The audio stream is processed within a Docker container and all generated data remains on the user's local computer. However, the user must ensure that he has the authority to store and process the file, for example if it contains personal data or other sensitive, GDPR-relevant information.