Touch Gesture Image Search via Semantic Context Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face difficulties in efficiently searching and identifying specific images on handheld devices due to small display interfaces and the need to scroll through numerous thumbnails, making it hard to find target photos quickly.
Innovation Solution
A related image searching method that allows users to select a context-of-interest area on a displayed image via touch input, analyze content characteristics, and search a database using algorithms like face, object, or scene recognition to identify matching images, displaying relevant files efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If users manually scroll through thumbnails to search for target photos, then they can find specific images, but the searching time and operational complexity increase significantly
Solution Approach 1:
The system performs preliminary analysis of image content characteristics (faces, objects, scenes) and organizes them into searchable databases before user search operations. This allows rapid retrieval based on semantic content rather than manual thumbnail browsing.
Solution Approach 2:
The patent introduces content characteristic analysis as an intermediary between the user and the image database. Instead of directly browsing thumbnails, users interact with semantic descriptors (face recognition results, object detection, scene classification) that mediate the search process and enable faster, more accurate image identification.
2Volume of moving object
If display interface size is reduced for compact handheld devices, then device portability improves, but image identification efficiency deteriorates
Solution Approach 1:
The patent transitions from two-dimensional thumbnail browsing to multi-dimensional semantic search space. Images are organized and searched based on multiple content characteristics (faces, objects, scenes, locations, time) simultaneously, enabling efficient retrieval on small screens by navigating through semantic dimensions rather than visual thumbnails.
Solution Approach 2:
The system changes the search parameters from visual display-based (thumbnail size, screen resolution) to content-based parameters (face recognition confidence, object detection accuracy, scene classification categories). This allows efficient searching independent of display size by operating in the parameter space of image content characteristics.
3Measurement precision
If comprehensive image analysis is performed to improve search accuracy, then search relevance improves, but processing complexity and energy consumption increase
Solution Approach 1:
The patent segments the image analysis process into distinct functional modules: face recognition, object detection, scene classification, location identification, and time extraction. Each module processes specific content characteristics independently, allowing selective execution based on user needs and reducing overall processing complexity while maintaining comprehensive search capability.
Data Source
AI summary
A related image searching method suitable for searching image files within a database is disclosed. The related image searching method includes steps of: selecting a context-of-interest from a displayed image according to a touch input event; analyzing a content characteristic in the context-of-interest area; determining an implication attribute of the content characteristic; and, searching the database according to the content characteristic and the implication attribute, so as to find out at least one image file with the same content characteristic or the same implication attribute. In addition, a user interface controlling method is also disclosed herein.