Key Posture Abstraction Using Entropy and Probability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current human posture analysis systems face challenges in efficiently and automatically abstracting key postures from digitized images due to high dimensionality and complexity, particularly in handling spatial-temporal information and lack of effective automatic methods for key posture selection and comparison.
Innovation Solution
The method involves abstracting spatial features using probability calculation, detecting key postures through entropy calculation, removing redundant postures, matching with templates in a codebook, and encoding the selected postures to enhance processing efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 3-D human model and articulated motion analysis are used, then posture recognition accuracy is improved, but computational complexity and processing cost increase significantly
Solution Approach 1:
The patent segments the human body into multiple key points (joints, landmarks) and processes them independently through detection and tracking modules. This divides the complex 3-D posture analysis into manageable components, reducing computational complexity while maintaining recognition accuracy through coordinated processing of individual key points
Solution Approach 2:
The patent introduces an intermediary representation system using 2-D projections and silhouette analysis as intermediate steps between raw image data and final 3-D posture recognition. This intermediary layer simplifies the computational burden by working with reduced-dimensional data before reconstructing full posture information
2Measurement precision
If comprehensive spatial-temporal information is captured from all video frames, then behavior characterization accuracy is improved, but processing time and data volume increase
Solution Approach 1:
The patent extracts only the essential spatial-temporal features from video frames by focusing on key point trajectories and motion patterns rather than processing all pixel data. This selective extraction maintains behavior characterization accuracy while dramatically reducing processing time through targeted feature analysis
Solution Approach 2:
The patent performs preliminary processing by detecting and tracking key points in advance before conducting full posture analysis. This preliminary action of identifying significant motion elements early in the processing pipeline reduces the computational load for subsequent behavior characterization while preserving essential temporal information
3Productivity
If automatic key posture selection is implemented, then system efficiency is improved, but algorithm complexity increases
Solution Approach 1:
The patent implements self-service through automatic key posture selection algorithms that autonomously identify and select significant postures without manual intervention. The system automatically determines which postures are key based on motion analysis and temporal patterns, improving efficiency while the algorithmic complexity is managed through rule-based decision making and heuristics
Data Source
AI summary
The method for automatic key posture information abstraction of this invention comprises the steps of: Abstracting from a series of continuous digitized images spatial features of objects contained in said images; abstracting shape features of said objects using a method of probability calculation; detecting key posture information contained in said series of continuous images using a method of entropy calculation; removing redundant key postures; mating obtained key postures with key posture templates stored in a codebook; and encoding mated key postures.