Visual Tagging for Object Association Across Multiple Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face difficulties in distinguishing and associating the same objects captured in different image data items with wide angles of view, as existing methods either superimpose or delete objects, leading to unclear associations between objects in multiple image data items.
Innovation Solution
A computer program that reads image data and object information, allows users to select and specify objects, and displays associated positions of objects with the same identification information across multiple image data items, using tags to connect objects visually, facilitating easy identification and association.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If multiple image data items with wide angle of view are displayed together with object information, then users can view objects in different images, but users face difficulties in distinguishing and associating the same objects across different image data items
Solution Approach 1:
The patent applies visual tagging where identical objects across different image data items are marked with the same visual tag (such as identical color coding or label). This allows users to easily distinguish and associate the same objects by their visual tags, transforming the abstract problem of object association into a concrete visual recognition task that resolves the contradiction between preserving object association information and ease of object identification
Solution Approach 2:
The patent introduces visual tags as an intermediary element that mediates between multiple image data items and user perception. These tags serve as visual cues that bridge the gap between different images, allowing users to associate objects across images without directly comparing the images themselves, thus improving ease of operation while maintaining object association information
2Device complexity
If existing methods superimpose or delete objects to combine image data, then image processing is simplified, but clear associations between objects in multiple image data items are lost
Solution Approach 1:
The patent segments the object association information from the image data itself by introducing separate visual tags. Instead of modifying the original image data through superimposition or deletion, the system adds independent visual markers that indicate object associations. This segmentation allows the image processing to remain simple while preserving object association information in a separate, non-intrusive layer
Solution Approach 2:
The patent adds a new dimension to the image display by incorporating visual tags as an additional layer of information. Rather than manipulating objects in the existing image dimension through superimposition or deletion, the system introduces visual tagging as a separate dimension of object identification, allowing association information to be preserved without complicating the original image processing
Data Source
AI summary
A computer program product for being executed on a computer causes an information processing apparatus to execute a process including reading an image data item from an image storage unit, and reading object information from an object information storage unit in which the object information relating to an object shown in the image data item is stored in association with identification information for identifying the object; displaying, by a display device, the image data items that have been read and the object information; accepting a selection of the object based on the object information, and accepting a specification of a position of the object with respect to the image data item; and displaying, in association with each other, the specified positions of the objects that have the same identification information, among the objects shown in different image data items for which the specifications of the positions have been accepted.