Adaptive grid generation for improved caching and image classification

摘要

Example embodiments of the present disclosure provide mechanisms for generating an improved grid for use in caching and classification when transmitting graphics data to a client computer. Such a grid may have the property that it is aligned to captured screen content such as user interface elements and borders between images and text. In some embodiments, an algorithm may be used that analyzes a screen and generates a subdivision grid of the screen that is likely to partition the screen into areas that substantially separate image and text content. This subdivision grid can also be further sub-divided (i.e., tiled) into regions that may be better suited for caching. In further embodiments, the algorithm may produce the grid using information from a window manager regarding window hierarchies and user interface element placement.