ComputerVision

    Keywords: Software

"Computer Vision" generally refers to the whole set of study, algorithms, and actual hardware and software needed to get a machine to usefully interpret visual data.

For go, this usually means being able to recognize a go board and stones, in spite of perspective, uneven lighting, obscured parts of the board, and sometimes intervening objects like people's hands, and then to produce data (typically SGF) representing a board position, or even a record of a whole game. Data usually comes from still images, and some projects aim to string the logic together to record a game from live or recorded video (sometimes considered an easier problem, since this starts from an empty board).

Programs and libraries that are specific to computer vision for go are:


This is a copy of the living page "ComputerVision" at Sensei's Library.
(OC) 2012 the Authors, published under the OpenContent License V1.0.
[Welcome to Sensei's Library!]
StartingPoints
ReferenceSection
About