Idea 1: FMS

2nd: Speech recognition

I am considering expanding the idea on week2's Cryptomnesia.

(I got mixed up with the term and the project relates more with False Memory Syndrome(FMS) or Ripley syndrome.)

The machine will keep overhearing my daily conversation with others and try to catch few keywords that contain keywords related to a childhood memory.

As humans imagine and visualize while making other's stories had their own story, the machine also creates an image to make the story as one's memory.

Flow

AI catch word in 'keyword pool'

⬇️

Gather the words that one heard along with the keyword

⬇️

Start creating an image based on the words and its sentiment

How to A) colab.research.google's libary that creates images that mataches a textual prompt

How to B) Get image from Unsplash API

How to C) Any other ML image API?

⬇️

Store the image in google photo with the timestamp on it

Step 1) Create a webpage that looks like a google photo

Step 2) if using google's colab, put random timestamp/ if using Unsplash API, use the data in it

⬇️

Caught by the human (the folder and the photo are caught by the user)

Things to think about