3D Display Depth Perception for Search Result Relevance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image segmentation and object recognition techniques fail to effectively search for and present objects of interest in video sequences when the user is unsure of the object's media asset or exact appearance, and struggle to display results in a meaningful manner.
Innovation Solution
The method employs depth perception in a three-dimensional (3D) display or virtual reality (VR) space to indicate search results and user preferences by generating relevance data for segmented and identified objects, displaying them in multiple planes based on user interest, with favored objects appearing in foreground planes and disfavored objects in background planes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image segmentation and object recognition techniques are used to search for objects in video sequences, then objects can be identified and tracked, but the techniques fail when the user is unsure of the object's media asset or exact appearance
Solution Approach 1:
The video sequence is divided into multiple clips, and each clip is further segmented into individual objects. This hierarchical segmentation allows the system to present multiple potential matches at different levels of granularity, enabling users to search without knowing the exact appearance or location of the target object.
Solution Approach 2:
The patent introduces a spatial dimension to the search results by displaying them in a three-dimensional space arranged according to their temporal occurrence in the video sequence. This 3D visualization provides users with an intuitive spatial map of when and where objects appear, compensating for the lack of precise appearance knowledge.
2Quantity of substance
If search results are presented in traditional two-dimensional displays, then all results can be shown, but users cannot quickly identify objects of interest among many results
Solution Approach 1:
Search results are displayed in a three-dimensional space rather than a flat 2D interface. Objects are positioned along temporal, spatial, and relevance dimensions, allowing users to quickly identify objects of interest by their position in 3D space. The depth perception provided by the 3D display enables users to prioritize certain results based on their visual prominence.
Solution Approach 2:
Different regions of the 3D display are assigned different meanings and visual properties. Objects closer to the foreground or center of the 3D space indicate higher relevance or more recent occurrence, while background objects represent less relevant or earlier occurrences. This local differentiation helps users quickly identify objects of interest without examining all results equally.
3Quantity of substance
If all search results are displayed with equal visual prominence, then completeness is maintained, but users cannot quickly distinguish objects of interest from non-interest objects
Solution Approach 1:
The 3D display creates visual hierarchy through depth and position, where objects of interest naturally appear more prominent in the foreground or central regions. This differential visual treatment preserves all search results while simultaneously encoding relevance information through spatial positioning, allowing users to quickly identify important objects without losing completeness.
Solution Approach 2:
The system can apply visual variations such as color, brightness, or size differences to objects based on their relevance to the user's search criteria. Objects that better match the search query or are more recently occurring can be highlighted or positioned to stand out, while maintaining the complete set of results in the background.
Data Source
AI summary
A method and apparatus are provided. The method provides display of preference information using depth perception, such as on a three dimensional display or in a virtual reality space. Objects are identified in an image or image sequence and assigned relevance information based on the preference information. Objects that are favored are shown in foreground planes of the display, to varying degrees based on the strength of the preference. Objects that are disfavored are shown in background planes of the display, also to varying degrees based on the strength of the preferences. Unidentified or neutral objects are shown at a base level, neither foreground nor background. An exemplary embodiment is provided for a movie database application with various actors shown pushed in or out. Another embodiment allows a user to adjust the plane of the objects to alter his preferences.