Web Page Hidden Content Detection via Layout Engine Comparison

The patent introduces a layout engine as an intermediary component between the web page and the detection system. The layout engine renders the web page and generates a render tree that reflects the actual visual layout, including hidden elements. This intermediary allows the system to detect hidden content that static text analysis cannot find, while maintaining a relatively simple detection process by comparing the DOM tree with the render tree.

The patent extracts the JavaScript execution function from the detection system itself and delegates it to the layout engine, which already has built-in JavaScript execution capabilities for rendering purposes. By separating the detection logic from the rendering logic, the system achieves high detection accuracy without significantly increasing the complexity of the detection apparatus, as the layout engine handles both rendering and detection tasks.

The patent transitions from a one-dimensional text-based comparison approach to a two-dimensional structural comparison approach by introducing the render tree. The render tree captures not only the HTML structure but also the CSS-styled visual layout information. By comparing the DOM tree with the render tree, the system can detect hidden content created through deformation HTML+CSS modes while maintaining relatively fast detection speed through efficient tree structure comparison algorithms.