Dynamic Weak Classifier Selection for Embedded Object Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for extracting specific objects from image data using cascade-connected weak classifiers face challenges in flexibly adjusting the tradeoff between extraction accuracy and process speed, particularly in embedded devices with varying specifications and operational conditions, requiring multiple classifiers and resources that are not efficiently managed.
Innovation Solution
An information processing apparatus with cascade-connected weak classifiers, a storage unit for processing content, a selection unit that determines which classifiers to use based on prerequisite conditions, and an extraction unit that employs evaluation values to select classifiers for flexible adjustment of extraction speed and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple cascade-connected weak classifiers are used to improve extraction accuracy, then extraction accuracy is improved, but device complexity and resource requirements increase
Solution Approach 1:
The patent implements dynamic selection of weak classifiers based on operational conditions. The system includes a selection unit that determines which weak classifiers to activate depending on device specifications, operational clocks, and use conditions. This dynamic configuration allows the same extraction unit to adapt its complexity level, using more classifiers when high accuracy is needed and fewer classifiers when resource constraints exist, thereby resolving the contradiction between accuracy and complexity.
2Measurement precision
If more weak classifiers are deployed to improve extraction accuracy, then extraction accuracy is improved, but processing time increases
Solution Approach 1:
The system dynamically adjusts the number of weak classifiers processed based on operational requirements. The selection unit can choose to activate only a subset of available weak classifiers depending on the desired tradeoff between accuracy and speed. When fast processing is needed, fewer classifiers are activated; when high accuracy is prioritized, more classifiers are activated. This dynamic control resolves the time-accuracy contradiction.
Solution Approach 2:
The patent changes the parameter of classifier activation status from static to dynamic. By modifying which classifiers are active based on operational conditions (such as operational clock speed, device specifications, and application requirements), the system can adjust processing time while maintaining acceptable accuracy levels. This parameter change enables flexible tradeoffs between processing speed and extraction accuracy.
3Adaptability or versatility
If multiple classifier configurations are prepared for different device specifications, then adaptability is improved, but device complexity and resource requirements increase
Solution Approach 1:
The patent creates a universal classifier management system that handles multiple device specifications through a single unified architecture. The selection unit serves multiple functions: it evaluates device specifications, determines appropriate operational parameters, and selects suitable weak classifier subsets. This multi-functional approach eliminates the need for separate classifier configurations for each device type, as the same system universally adapts to different specifications, thereby improving adaptability without proportionally increasing complexity.
Data Source
AI summary
In an information processing apparatus that processes data using cascade-connected weak classifiers, processing specification information specifying the processing content of each of the weak classifiers is stored. The weak classifiers to be used in processing the data are selected from the weak classifiers by referring to a table in which is specified information for determining the weak classifiers to be used based on a condition for processing the data. The data is then processed by the selected weak classifiers based on the processing specification information that corresponds to those weak classifiers, and an object is extracted from the data using an obtained evaluation value. Through this, a combination of extraction process speed and extraction accuracy can be changed in a flexible manner when extracting a specific object from image data.