Trivial Pursuit OCR Digitization
Budget: $30 – $250 USD
I have a complete Trivial Pursuit deck—400 brand-new, perfectly clean cards, each with six Q&A pairs laid out in identical sections. I will ship the entire game to you (and you’re welcome to keep it once the work is finished).
Your task is straightforward: scan every card, run accurate OCR, and return a single JSON file that looks like this:
[
{"q": "Question 1", "a": "Answer 1"},
{"q": "Question 2", "a": "Answer 2"},
… 2400 objects total …
]
Because the layout never changes, a template-based approach should make extraction fast and highly accurate. Feel free to use Tesseract, Google Vision, ABBYY, or any Python/OpenCV pipeline you prefer; what matters to me is clean, verified text with matching answers.
Acceptance criteria
• One validated JSON file containing exactly 2 400 objects, UTF-8 encoded
• Spot-checked accuracy ≥ 99 % on both questions and answers
• The game will be in French
Once you confirm the JSON meets the spec, the project is done—and the game stays with you.
Your task is straightforward: scan every card, run accurate OCR, and return a single JSON file that looks like this:
[
{"q": "Question 1", "a": "Answer 1"},
{"q": "Question 2", "a": "Answer 2"},
… 2400 objects total …
]
Because the layout never changes, a template-based approach should make extraction fast and highly accurate. Feel free to use Tesseract, Google Vision, ABBYY, or any Python/OpenCV pipeline you prefer; what matters to me is clean, verified text with matching answers.
Acceptance criteria
• One validated JSON file containing exactly 2 400 objects, UTF-8 encoded
• Spot-checked accuracy ≥ 99 % on both questions and answers
• The game will be in French
Once you confirm the JSON meets the spec, the project is done—and the game stays with you.