Visual Search Linking Real-World Objects to Virtual Data

The patent replaces the mechanical keyboard input system with an optical recognition system. Instead of physically pressing multiple keys to input characters, the system captures an image containing text, performs optical character recognition (OCR) to convert the visual text into searchable text data, and automatically extracts search keywords. This substitution eliminates the need for manual character entry while maintaining the information retrieval function.

The patent performs preliminary text extraction and keyword identification from the captured image before the user initiates the search. The system automatically processes the image, extracts visible text using OCR, identifies relevant keywords, and prepares the search query in advance. This preliminary action eliminates the need for users to manually enter search terms, thereby improving both content retrieval efficiency and user interface efficiency.

The patent enables the content itself to provide searchability information without requiring external metadata insertion. The system extracts text directly from the image content using OCR, allowing the visual content to serve its own search function. This self-service approach eliminates the time-consuming metadata insertion process while maintaining reliable search results, as the text is extracted directly from the source material rather than from potentially incomplete or inaccurate metadata.