Explainers - doraithodla/notes GitHub Wiki

#Idea - From Text to Explainer video

  1. User proovides a script
  2. The ChatBot analyzes the script and identifiels characters, scenes, animations
  3. The ChatBot creates an Explainer Video along with audio

Example:

"A man and a dog walk along in a park"

ma, dog

park (animation) background

man and dog walking (animation) foreground

ChangeLog

  • 5//29: Talked to Ananya to create some animations
  • 5/30: Moving back to RA and discontinuing PyTutor since it does not provide major advantages over ChatGPT4 now
  • 5/30: RA topic Liver diseases (there seems to be so little consolidated information about it)
  • 5/30: Secondary long term project is Explainer Maker

Explainer Video Co-pilot

What does it do? From a script(prompt) generates an explainer video.

Why? Makes it easy for students to learn from Explainer videos Explainer videos are hard to produce (cost 5oK to 2 lakhs) even for short one minute videos. It takes a lot of time and requires animators (which are hired by game companies)

How Do we do it? User writes a script The director specifies a character set (equivalent to the cast in a movie) The co-pilot reads the script (prompt) and generates scenes (video) and audio. The system uses stable diffusion or a similar tool to generate video streams

Marketplace Education (on YouTube) Schools/Colleges Integrated it to tutor assistants. Why? Lot of video content is being consumed and predicted to be consumed. https://www.globenewswire.com/news-release/2023/04/27/2656791/0/en/Rapid-Growth-of-Large-Language-Models-Drives-Generative-AI-Market-to-Projected-51-8-Billion-by-2028.html