S.M.A.R.T. is a Summarization Model for Annotation Recognition & Translation/Transcription that performs quick, low error and detailed extraction of summary, action items, sentiment analysis,name entity recognition, timeline extraction, creates recommendations, translates these into different languages, creates transcription from hebrew, farsi and english lanaguages which then translates the full transcription in the other two languages if one of these is the language the audio/video file is in. It processes audio and video file using ffmpeg in seconds converting it into .wav format that is then processed for the above operations.
- Summarization
- Action Items Recognition
- Sentiment Analysis
- Timeline Extraction
- Name Entity Recognition
- Creates Recommendation
- Translates
- Transcribes
- Converts audio/video into higher quality .wav
- Has a strong UI system with ease of operation
- Gives updates on current process
- Shows the time taken for most operations
- Breaksdown audio into chunks for effective translation
- Runs with no cloud dependency or Network access
- Suitable for CPU operations
- Efficient and quick in the above operations
Edge computing for privacy and quick meeting brief in places were data security is vital and network/cloud issues are persistent
Download the installer from Google Drive and download entire folder then open the SMART_setup.exe file and installation will be automatic.
Please follow the video from Youtube for a demo on how to use the software
Was originally made as part of IWTSD (Department of Defense) Smart Summary challenge on Innocentive platform