Object Classifier Hierarchy for Dynamic Scene Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods are computationally intensive for detecting characteristics in dynamic scenes, necessitating more efficient approaches for object classification in sequential image data frames.
Innovation Solution
An image processing system comprising an image data interface, an object classifier, and storage for categorization data organized in an object definition hierarchy, allowing selective execution of subsets of object definitions in object classification cycles to optimize processing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all object definitions are executed in every object classification cycle, then comprehensive object detection accuracy is improved, but computational complexity and processing time increase significantly
Solution Approach 1:
The patent segments the complete set of object definitions into multiple subsets and divides object classification cycles into different types (first type cycles execute first subset, second type cycles execute second subset). This segmentation allows the system to balance between comprehensive detection and processing efficiency by selectively executing different subsets in different cycles.
Solution Approach 2:
The patent implements periodic action by alternating between first type object classification cycles and second type object classification cycles. Each cycle type executes a different subset of object definitions, creating a periodic pattern that maintains detection accuracy while reducing average computational load compared to executing all definitions in every cycle.
2Reliability
If all object definitions are executed in every object classification cycle, then complete object classification coverage is improved, but computational resource consumption increases
Solution Approach 1:
The patent segments object definitions into multiple subsets (first subset and second subset) and assigns them to different cycle types. This segmentation enables the system to maintain classification coverage by executing different subsets across different cycles while reducing instantaneous computational resource consumption in any single cycle.
Solution Approach 2:
By periodically alternating between first type cycles (executing first subset) and second type cycles (executing second subset), the system distributes computational resource consumption over time, preventing peak resource demands while ensuring that both subsets are executed periodically to maintain complete classification coverage.
3Productivity
If a single subset of object definitions is used, then processing efficiency is improved, but adaptability to different object classification needs deteriorates
Solution Approach 1:
The patent makes the object classification system universal by implementing multiple subsets of object definitions and multiple cycle types. The system can adapt to different classification needs by selecting appropriate cycle types and subsets, providing multi-functionality that handles both speed-critical scenarios and comprehensive detection scenarios.
Solution Approach 2:
The patent introduces dynamics by allowing the system to change the executed subset of object definitions based on the cycle type. First type cycles dynamically execute the first subset while second type cycles dynamically execute the second subset, enabling the system to adapt its behavior to different processing requirements and maintain both efficiency and versatility.
Data Source
AI summary
An object classifier using a set of object definitions arranged in an object definition hierarchy including at least a first group of coarse-level object definitions and a second group of finer-level object definitions. The object classifier is arranged to configure a first object classification cycle and a second, subsequent, object classification cycle by selectively executing a first subset of object definitions from the categorization data in the first object classification cycle; and selectively executing a second, different, subset in the second object classification cycle.