A biology-based template-matching framework

Abstract

The multichannel model is a complete reassessment of how neurons work at the biochemical level. Its results can be extended into an overarching theory of how vision, memory and cognition come to be in the living brain. This article documents a first attempt at testing the model's validity, by applying its principles to the construction of an image template-matching framework, which is then used to solve a Graphical User Interface (GUI) automation problem. It is found that the template-matching application thus implemented can consistently locate required visual controls, even when template and match differ by color palette or (to a slighter degree) feature proportion. The article concludes by discussing the significance of these results and directions for further research.