UI Element Detection Using Combined Selector, CV, and OCR Attributes

The patent combines multiple graphical element detection techniques (selectors, computer vision, OCR) into a unified detection system. The system merges these techniques by collecting UI descriptor attributes from all techniques and selecting the most accurate match based on predefined accuracy thresholds, thereby achieving higher detection accuracy without requiring complex individual techniques to function independently

The invention creates a universal detection framework that can handle multiple UI element types and scenarios using a single integrated approach. The system universally applies multiple detection techniques to various UI elements and automatically selects the appropriate technique or combination based on the element's characteristics and the detection accuracy requirements

The patent merges multiple detection techniques into a single coordinated system that collects attributes from selectors, computer vision, and OCR techniques. The system then processes these attributes together and selects the most accurate match based on accuracy thresholds, achieving high precision for diverse scenarios while managing complexity through unified processing

The invention segments the detection process into distinct phases: collecting UI descriptor attributes from multiple techniques, processing and comparing these attributes, and selecting the final match based on accuracy thresholds. This segmentation allows each technique to contribute its strengths while the overall system maintains manageable complexity through structured processing