Transcribing Audio from Family Stories and Building a Graph
Chinese New Year is when relatives get together. Stories can get shared which go back a few generations. AI could be used to:
1) Detect the persons telling the story. It might be two aunties sharing a story about their parents. Need to determine that there are two people telling a story and that they are talking about their relatives.
2) For each person, determine their role and relationships. One auntie might be talking about her father. Her father may have had 2 wives. Those wives have children. That auntie is a child from the first wife.
The other auntie might be your mother in law, and both aunties are cousins to each other. Your mother in laws father had a younger sister, and that sister had a daughter. That daughter is the other auntie.
3 The system needs to determine who is telling the story, they would be marked as primary sources for this story. The people mentioned in the story also have to be detected. Then based on the story content, relationships can be graphed.
4 For fun, the system can guess what the aunties look like based on their voice. It could even propose some guesses based on the faces it already has from photos you’ve uploaded.
