Subject Detection Discriminator Selection for Varying Lighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional subject detection devices face challenges in accurately detecting face regions in images, especially under varying photographing environments such as backlighting and low illumination, where existing technologies struggle to maintain detection accuracy.
Innovation Solution
A subject detection device and method that sequentially acquires images and selects a specific discriminator from a group corresponding to different photographing environments, employing a selection unit to choose the appropriate discriminator for accurate face detection, switching between discriminators if necessary to improve detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single discriminator is used for subject detection, then the device complexity is reduced, but the detection precision deteriorates under varying photographing environments
Solution Approach 1:
The patent divides the detection system into multiple discriminators, each specialized for specific photographing environments (e.g., normal lighting, backlighting, low illumination). The selection unit segments the overall detection task by choosing the appropriate discriminator based on environment type, allowing high precision in each segment while managing complexity through organized classification of discriminators by environment category
Solution Approach 2:
The system changes the parameter of discriminator selection based on photographing environment parameters. The selection unit adjusts which discriminator is active by evaluating environment characteristics (lighting conditions, contrast levels) and switching between pre-trained discriminators optimized for different parameter ranges, thereby maintaining high detection precision across varying conditions
2Measurement precision
If multiple discriminators are prepared for different environments, then the detection precision improves, but the device complexity increases
Solution Approach 1:
Multiple discriminators are pre-trained and prepared in advance for different photographing environments before actual detection begins. Each discriminator is pre-optimized for specific conditions (backlighting, low light, normal lighting), so when detection is needed, the system simply selects from these ready-made options rather than adapting a single discriminator in real-time, improving both precision and efficiency
Solution Approach 2:
The selection unit serves as a universal control mechanism that manages multiple specialized discriminators. Rather than having separate detection systems for each environment, one selection unit universally handles environment assessment and discriminator selection, allowing the system to maintain multiple specialized components while managing them through a single coordinating element
3Adaptability or versatility
If environment-specific discriminators are used, then the adaptability to different photographing environments improves, but the processing time increases due to discriminator selection
Solution Approach 1:
Discriminators are pre-trained and categorized by environment type in advance. The selection unit has pre-established criteria for matching environment characteristics to appropriate discriminators. This preliminary preparation eliminates the need for complex real-time optimization, allowing rapid environment assessment and discriminator selection that minimizes processing time while maintaining high adaptability
Solution Approach 2:
The selection unit continuously monitors photographing environment characteristics and provides feedback to determine the most appropriate discriminator. This feedback mechanism allows the system to adapt to changing conditions in real-time by selecting discriminators based on current environment parameters, achieving high adaptability without requiring exhaustive processing of all possible environments
Data Source
AI summary
An image capture device (1) includes: a storage unit (19) that stores a plurality of discriminators corresponding to a plurality of photographing environment types, respectively; an acquisition unit (42) that sequentially acquires images; a selection unit (41) that selects a specific discriminator from among a plurality of discriminators that corresponds to the plurality of photographing environments, respectively; and a detection unit (43) that detects an image of a subject in images sequentially acquired by the acquisition unit.