Page 744 - AI for Good Innovate for Impact
P. 744
AI for Good Innovate for Impact
speech production. Through nationwide deployment via open-source system access and
optimized teaching protocols, this initiative will empower the deaf community to achieve
lifelong learning and personal development.
Through continuous technological iterations and public benefit initiatives, we are dedicated to
building a more inclusive and equitable global education ecosystem for the hearing-impaired
community, fully realizing the SDG principle of “leaving no one behind.”
System Maintenance:
To enhance system reliability and user experience, we improve functional stability by addressing
user feedback, refine lip-reading algorithms through analyzing real-world bad cases to boost
accuracy, and conduct regular inspections of model deployments and server operations for
sustained performance.
3 Use Case Requirements
REQ-001: Users must possess network-enabled terminal devices equipped with video capture
capabilities.
REQ-002: The cloud-based deployment of the system requires sufficient computational power
to support concurrent user access.
REQ-003: The core functionality of lip-reading recognition lies in converting user captured
video footage into textual recognition outputs.
REQ-004: The AI system must efficiently and effectively organize, utilize, and expand training
and testing datasets.
4 Sequence Diagram
708
