Skip to main content

Features and Requirements

Functional Requirements

Voice Recognition

  • Voice recognition should be able to listen and understand users.
  • Voice recognition may convert speech to text

Mood Detection

  • Application should be able to detect mood of child based on voice
  • Application should have fallback options if it detects child is distracted.

Quiz Fallback

  • If the user answers a quiz question(s) wrong, there should be fall back options.
  • Fall back options may include multiple choice questions, replaying a part of the video , or question layering.
    • Application must ask AI to generate multiple choice questions and question Scaffolding
    • Humans must review questions before they are available to users.
  • Additional spotlight mode fallback. Video will rewind and a spotlight will shine to show where correct answer is.

No Distraction Mode

  • Parent account should be able to toggle No Distraction Mode for videos on the kids account.
  • No Distraction Mode should:
    • turn off hyperlinks for the video
    • restrict fast forwarding
    • restrict rewinding
    • not allow jumping around the video.

Question Scaffolding

  • If child answers question wrong then AI will generate additional questions.
  • The questions will slowly lead child to correct answer to original question, after responses.

User Data

  • User data should be saved and securely logged in the database (user id, video id, watch time, correct / incorrect answer)
  • There should be a dashboard that contains:
    • Summary
    • Users
    • Settings

Mascot

  • Mascot must be overlayed over videos
  • Mascot must 'read' the questions and answers outloud.

Account

  • Users must be able to create their own accounts (username, password, accountID)
  • Users must have different permission levels that will give them access to which videos they can watch. Think parental controls: the parent’s account can control what the kids account can / cannot watch.

Nonfunctional Requirements

  • Improvements to voice recognition:
    • Voice recognition needs to be able to interpret less ‘perfectly worded’ input (ie the screaming of a child, different accents, etc)
    • Should be able to map input to answers, even if they are not the same
  • Response time needs to be improved on
  • No voice data will be stored, in order to comply with the law (children voice)
  • All data that may be stored shall be secure and private, only available to the kid account and the associated parent account.
  • Improve AI response latency speed by 30%
  • AI should able to generate good questions for Question Scaffolding quickly and on the fly.
  • Data should be saved within a reasonable time frame.
  • Data retrieval from database should be fast.
  • Application should accurately detect mood from the two options 'Distracted' and 'Not Distracted' with a 5% margin of error.
  • No Distraction Mode should accomplish it's goals of keeping the child on track and shouldn't be easily bypassed.