Multimedia Image Processing for Emotional Interaction Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing monitoring systems for children fail to select images that capture special expressions, movements, and interactions between characters, resulting in a lack of richness and narrative content.
Innovation Solution
A multimedia image processing method that uses artificial intelligence to identify and select images based on specific conditions such as emotional body movements, similar directions, and interactions between characters, producing a concatenated video with richer thematic content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If AI-based selection focuses only on composition, then selection simplicity is maintained, but image content richness deteriorates
Solution Approach 1:
The patent segments the image selection process into multiple independent analysis dimensions: composition evaluation, expression recognition, movement detection, and interaction analysis. Each dimension is processed separately by dedicated modules, allowing comprehensive evaluation without complicating the overall system architecture. This segmentation enables the system to maintain operational simplicity while capturing rich image content across multiple criteria.
2Loss of information
If multiple selection criteria are added to capture expressions and interactions, then image content richness is improved, but system complexity increases
Solution Approach 1:
The patent implements a universal processing framework where a single image input is simultaneously analyzed across multiple functions: composition assessment, facial expression recognition, body movement detection, and character interaction analysis. The system uses multi-functional modules that can handle different analysis tasks within a unified architecture, reducing overall system complexity while achieving comprehensive image evaluation.
Solution Approach 2:
The patent introduces intermediary processing layers that bridge raw image data and final selection decisions. These intermediaries include feature extraction modules that convert raw pixels into meaningful attributes (expressions, movements, positions), and integration modules that synthesize results from multiple analysis dimensions. This intermediary structure manages complexity by organizing the processing pipeline into manageable stages.
3Speed
If AI only judges composition, then processing speed is maintained, but selection accuracy for vivid images deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-defining selection criteria and thresholds for expressions, movements, and interactions before actual image processing begins. The system pre-configures what constitutes a noteworthy expression or interaction pattern, allowing rapid comparison against captured images without complex real-time decision-making. This preparation enables fast processing while maintaining high selection accuracy for vivid moments.
Data Source
AI summary
A method for multimedia image processing includes steps of identifying objects and selecting images, and is characterized in that when a body position of a preset object is detected and conformed to match a preset posture; or a plurality of the preset objects are detected to have body movements and facial expressions that are emotional, and the preset objects have at least one of looking in similar directions, one looking at the other, and at least two looking at each other, an interception time point is selected for selecting a candidate image, and the candidate image can be collected to produce a concatenated video with rich contents. An electronic device is also introduced for multimedia image processing, a terminal device connected thereto, and a non-transitory computer-readable recording medium.