Rapid workflow design using machine learning

Abstract

A system and method for rapid workflow design utilizing machine learning includes a scanner, a convolutional neural network, a converter, a graph generator, and a controller. The scanner may convert a paper copy of a process flow into a pixelated image. The convolutional neural network is configured to segment the pixelated image into objects including at least one of an action, a decision, a connector, or a combination thereof. The converter is configured to transform the objects into a connectivity matrix. The graph generator is configured to transform the connectivity matrix into a rectilinear graph. The controller is configured to identify automated actions, semi-automatic actions, and actions requiring operator input from graphical icons provided on the paper copy and, based on the rectilinear graph, to serialize the automated actions, the semi-automatic actions, and the actions requiring operator input as control commands to a printing system.

Dependencies are inferred from the claim text.