So far this method doesn't work:
If you want to work on this.
- Start from the center of a mobile phone screen.
while cursor is not at button position:
- overlay a cursor ontop of the image
- Ask GPT Left, Right, Up or Down
- ex. GPT says Left
- Go half way left
- overlay a cursor ontop of the image
- Ask GPT Left, Right, Up or Down
- ex. GPT Says Right
- Go half of the half way right
- position found!
but this one does:
simpler solution (void and null), split the image into sections. super impose a red grid over the image, and ask gpt which square the item is in. Possibly use a combination of both to get accurate results.