Information Processor Object Search via Pre-Indexed Metadata
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information processors for video search lack user convenience as they restrict object selection to the currently reproduced frame, preventing users from easily finding objects in other frames and do not allow dynamic adjustment of search conditions during video reproduction.
Innovation Solution
An information processor with a recording unit, reproduction unit, output unit, and control unit that allows users to specify object conditions and display occurrence information of non-reproduction objects, enabling flexible object selection and condition adjustment during video reproduction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If object selection is restricted to currently reproduced frames, then device complexity is reduced, but ease of operation deteriorates as users cannot easily find objects in other frames
Solution Approach 1:
The system performs preliminary indexing of all objects across the entire video data before reproduction begins. Object management information including occurrence time zones is extracted and stored in advance, enabling users to search and select objects from any frame without actually reviewing the entire video content manually.
Solution Approach 2:
An object management information structure serves as an intermediary between the video data and user search operations. This intermediate layer contains pre-extracted object information including names, types, and occurrence time zones, allowing users to efficiently search and select objects without direct frame-by-frame analysis.
2Adaptability or versatility
If search conditions are fixed before reproduction starts, then device complexity is reduced, but adaptability deteriorates as users cannot change conditions during reproduction
Solution Approach 1:
The search condition specification is made dynamic, allowing users to add, remove, or modify object conditions during video reproduction. The system continuously monitors reproduction progress and updates available search options based on current playback position and pre-indexed object data.
Solution Approach 2:
The system provides feedback to users during reproduction about objects that have occurred and are available for selection. This feedback mechanism enables users to make informed decisions about changing search conditions based on real-time reproduction status and previously extracted object information.
3Measurement precision
If users must manually review frames to find objects, then measurement precision is improved for object identification, but loss of time increases significantly
Solution Approach 1:
Object information extraction and indexing is performed in advance during a preliminary processing stage. All objects across the entire video are identified, classified, and their occurrence time zones are recorded before reproduction begins, eliminating the need for manual frame review during actual use.
Solution Approach 2:
Instead of requiring users to view and analyze actual video frames to identify objects, the system creates and uses copies of object information in the form of structured metadata including object names, types, and temporal occurrence patterns. This copied information enables rapid searching without time-consuming frame review.
Data Source
AI summary
When video data is reproduced by a recording/reproduction unit, an output unit outputs occurrence information on an object, which is not included in the video data being output by the output unit but is included the video data recorded on a recording medium, to a display unit. While the video data is being reproduced from the recording medium by the recording/reproduction unit, an instruction specifying a predetermined object condition including a non-reproduction object is input based on the occurrence information and the reproduction operation is performed by extracting video data satisfying the object condition.