News for hackers. Stuff that matters.
Re:How We Built the ARKit Sudoku Solver by bnjmn on Tuesday October 10, 2017 @10:34AM (#15442098)
Since the author spent so much time on the optical character recognition step, it's worth mentioning that you don't even really need OCR for this task.

You can just find the squares, group the characters in the squares into visual equivalence classes, assign each class an arbitrary number, solve the puzzle in terms of those numbers, then fill in each empty square with the (average?) image of the equivalence class it matches.

This would allow you to solve a Sudoku puzzle with letters or WingDings instead of numbers, and the output font would naturally match that of the original puzzle.

Re:How We Built the ARKit Sudoku Solver by spaceheeder on Tuesday October 10, 2017 @11:29AM (#15442611)
If the app still wanted to support puzzles that had notes scribbled in them then it would need some kind of OCR to tell the difference between a "starter" cell with known-good data and "puzzle" cells it needs to solve.

But treating the symbols in the starter cells as arbitrary is ingenious, imo!

Re:How We Built the ARKit Sudoku Solver by gok on Tuesday October 10, 2017 @11:29AM (#15442609)
Using OCR means you can solve a Sudoku with some printed and some hand-written numbers though.
Re:How We Built the ARKit Sudoku Solver by yeldarb on Tuesday October 10, 2017 @10:56AM (#15442298)
That's an interesting idea! What happens if the source puzzle doesn't have all of the digits present though?
Re:How We Built the ARKit Sudoku Solver by ndh2 on Tuesday October 10, 2017 @11:02AM (#15442359)
Take random characters from WingDings.
Re:How We Built the ARKit Sudoku Solver by bnjmn on Tuesday October 10, 2017 @03:10PM (#15444490)
This! Still, admittedly, this is a drawback of my strategy.