Rizky Yanuar Kristianto, Muhammad Abduh, Hasanuddin Al-Habib
The popular puzzle game Sudoku challenges players to fill a 9x9 grid with digits from 1 to 9 so that each row, column, and 3x3 sub-grid contains all the digits without repetition. This paper explores an integrated approach leveraging advanced image processing and artificial intelligence techniques to automate Sudoku solving. Initially, digit classification using convolutional neural networks (CNNs) identifies and distinguishes digits on the Sudoku board. The system accurately locates and isolates Sudoku squares within images by implementing contour detection and perspective transformation. Finally, a backtracking algorithm ensures efficient and precise placement of numbers to solve the puzzle. Experimental results demonstrate high accuracy in digit recognition and Sudoku solving, showcasing the effectiveness of integrating cutting-edge technologies to enhance problem-solving capabilities in logic-based puzzles. From this research, we obtained the average time computing of the system to recognize the sudoku puzzle and predict its solution, which was 0.0582. © 2025 Author(s).
Data Science Study Program, Universitas Negeri Surabaya, Surabaya, Indonesia
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock