Interactive homage to the Markov chain-based work of Hiroshi Kawano (1925-2012).
Drop an image on the canvas to pixelate it into a grid with a color palette taken from Kawano. A probabilistic Markov Random Field model is then created of the image from the von Neumann neighborhoods (with a Manhattan distance of 1) of all the cells in the 40 x 40 cell grid.
Kawano originally worked with Markov chain models of sequences of cells instead of their two-dimensional analogs, MRF, that are used here.
Cells in the grid are then replaced iteratively in accordance with the probabilities in the MRF model, in a process I call Markovian drift. That is, cells are allowed to be determined by their closest neighbors to the left, top, right and bottom.
Press [V] to replace cells from left to right, [H] to replace cells from top to bottom, or [R] to replace cells in random order.
Press [Space] or left click to pause. Press [Return] to reset to initial grid.
Open edition for 30 days.
