Multi-Focus Camera Apparatus for Robust Target Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting and protecting personal information in images, such as those using mosaic processing, face challenges in maintaining detection performance under changing photography conditions and subject movement, leading to incomplete protection of target objects.
Innovation Solution
A multi-focus camera apparatus and image processing method that utilize detection, region acquisition, and image processing means to identify and display target objects with consistent blurriness relationships across multiple images, applying techniques like region fill, mosaic, or template insertion processing to ensure whole target regions are differentiated from other areas, even when initial detection fails.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pattern recognition according to an artificial intelligence algorithm is used to detect target objects, then detection can be performed on target objects in images, but detection performance deteriorates due to changes in photography conditions and subject movement
Solution Approach 1:
The patent transitions from single-image pattern recognition to multi-image analysis by acquiring images at multiple focal lengths. This dimensional change from 2D single-frame analysis to multi-focal-length 3D space analysis enables more robust target detection that is less sensitive to photography condition changes and subject movement.
Solution Approach 2:
The system changes the focal length parameter of the camera to acquire multiple images with different depth of field characteristics. By varying this optical parameter, the system can identify target objects across different focus planes, making detection more reliable under changing conditions.
2Reliability
If mosaic processing is applied to protect personal information, then target objects can be obscured, but incomplete detection leads to failure to protect entire target objects
Solution Approach 1:
The system performs preliminary multi-focal-length image acquisition and integrated analysis before applying mosaic processing. By pre-identifying the complete target object region through multi-image correlation, the subsequent protection processing can be applied more reliably to ensure entire target objects are obscured.
Solution Approach 2:
The patent merges multiple images taken at different focal lengths to create a comprehensive view of the scene. This combination allows the system to detect target objects that may be out of focus in individual images, thereby improving both detection accuracy and protection reliability.
3Productivity
If single-image detection methods are used, then processing is simple and fast, but the whole target object region cannot be reliably identified under changing conditions
Solution Approach 1:
The patent segments the image acquisition process into multiple focal length captures, allowing parallel processing of individual images followed by integration. This segmentation enables the system to maintain processing efficiency while improving target identification reliability through multi-view analysis.
Data Source
AI summary
In a multi-focus camera apparatus that photographs the same region with cameras having different focal lengths, detection means for detecting a specified target in images obtained by the cameras, region acquisition means for obtaining in the images regions comprising pixels having substantially the same relationships between blurriness degrees as relationships between blurriness degrees of the specified target having been detected in the respective images, and image processing means for carrying out image processing to display the obtained regions differently from other regions are included.