Althea Wallop & Joel Swanson

For this speculative project, we utilized Optical Character Recognition (OCR) to analyze found connect-the-dots puzzles. Our objective was to catalog the positions of all the dots along with their corresponding numerical markers. This dataset was then used to generate animations of the puzzles being drawn autonomously, without human input. We found EasyOCR to provide the best results.

This project presented several interesting challenges, such as disambiguating between dots and other symbols within the puzzles, such as stars, and correctly pairing dots with their corresponding numbers.