Game Record

Steven Scher and Ryan Crabb

UCSC CMPS 290b, Fall 2007

 

MOTIVATION

Many board games, such as Chess and Go, are played both in-person and online. After playing a game online, the record of the game is available so that the players may discuss particularly good and bad moves with each other and with friends and teachers. This commentary often adds significantly to the social experience and skill level improvement. Very good players remember the games played in person, and review them afterwards. Other players don’t have as keen memory, but they do have a cellphone.

CHALLENGE

A players places their cellphone camera on the table next to the board during a game. Our software on the phone automatically take pictures, finds the board and detect moves. A complete record of moves played is created, so that good and bad moves can be reexamined by those present and shared via email with others. A file in the standard SGF file format may be uploaded to the player's online account, and online game databases can be queried to find others' games facing similar decisions.

CONTRIBUTION

We seek tremendous reliability, hoping to allow many games to be recorded without error, giving requirements of a maximum false negative rate when detecting stones of 1 in 1000, and false positive rate of 1 in 10000. To achieve this, we will apply camera calibration techniques to find the 3D position of the board and model each stone as a 3D object to create a customized stone detector for each location in the image. Additionally, identifying foreground objects such as hands in order to ignore them will be a major challenge.

PAPER:

Full Paper (Draft)

INTRODUCTION

RELATED WORK

FIGURES

REFERENCES

Recording Games of Go

* Automatic Extraction of Go Game Positions from Images: An Application of Machine Learning to Image Mining, Alexander K. Seewald

* Detection of Go-board contour in real image using genetic algorithm, Shiba, K.; Mori, K. SICE 2004 Annual Conference Volume 3, Issue , 4-6 Aug. 2004 Page(s): 2754 - 2759 vol. 3

* Recording a Game of Go: Hidden Markov Model Improves a Weak Classifier, Steven Scher and Manfred Warmuth, University of California Santa Cruz (My project from last year)

* GoCam: Extracting Go Game Positions from Photographs, Teemu Hirsimäki, Helsinki University of Technology.

* Image2SGF: Chris Ball's perl script (http://www.inference.phy.cam.ac.uk/cjb/image2sgf.html)

Calibration: Finding a Homography from detected Points and Lines

* Hartley & Zisserman: Multi View Geometry

* Metric Rectification for Perspective Images of Planes, David Liebowitz and Andrew Zisserman, University of Oxford

* A New Sub-Pixel Detector for Grid Target Points in Camera Calibration, Chen D., Wang Y. and Zhang G., Beijing University of Aeronautics and Astronautics

* Houghing the Hough: Peak Collection for Detection of Corners, Junctions, and Line Intersections, Barrett and Petersen, Brigham Young University

Detecting Stones

* Hough Transform

* Pattern Classification. Duda, R., Hart, P., and Stork, D. John Wiley and Sons, Inc, 2001.