Why doesn't Ranorex rely on image-based recognition by default?

Image-based recognition matches a saved picture of a control against the current screen, which makes it sensitive to almost any visual change - a different screen resolution, DPI scaling, operating-system theme, font rendering, or translated text can all cause a previously reliable image match to fail even though the underlying control hasn't actually changed.

Reading the real object model instead - names, roles, and attributes rather than pixels - is both more resilient to those cosmetic changes and considerably faster, since it doesn't require rendering and scanning a bitmap on every action.

Ranorex does still support image-based locators as a fallback, and they're genuinely useful for the narrower set of cases where no accessible object model is available - certain legacy applications, some virtualized or remote desktop sessions (like Citrix), or canvas-drawn UI - but they are treated as the exception rather than the default identification strategy.