Features and Requirements
Functional Requirements
- Audio input will be captured via device's microphone
- Whisper AI will convert the speech to text
- User will be able to toggle the microphone on/off through AAC Board
Virtual AAC Board
- System will be organized in a color key organization to make simpler access to similar tiles
- Categories will be a folder in which corresponding tiles will be placed
Context Analysis
- The system will analyze text to determine key context words
- The model will compare the key words to the AAC vocabulary list to determine best possible suggestions and show through opacity of each tile
Nonfunctional Requirements
- Accurate speech to text transcription
- Speech to text transcriptions should be low latency( ~ 3-4 second lag to transcribe what was said)
AAC Board Coherence
- Picture symbols must stay in the same place as to not interfere with the users muscle memory
- Picture symbols can be grouped by topic so that specific words can be easily found.