Inclined Reflector Image Segmentation for Multi-Angle Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information processing systems face challenges in accurately recognizing objects from a specific direction without adjusting the posture of the subject, particularly in environments like cash register terminals where multiple items need to be identified with high precision.
Innovation Solution
An information processing system that utilizes a camera with a reflecting unit having an inclined reflecting surface to divide an image into partial images, performs object recognition on each partial image, and evaluates recognition accuracy based on a reliability threshold, allowing for precise object identification from multiple angles without adjusting the object's posture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If object recognition is performed on a single taken image, then processing speed is improved, but recognition accuracy deteriorates when the object posture is not adjusted
Solution Approach 1:
The patent divides the taken image into multiple partial images based on a grid pattern, and performs object recognition on each partial image separately. This segmentation allows the system to identify objects even when they occupy only part of the image or when the object posture is not optimized for single-image recognition, thereby improving recognition accuracy without requiring multiple full-image captures.
Solution Approach 2:
The patent introduces a directional dimension by performing object recognition on multiple partial images from different regions of the taken image. By analyzing objects in multiple directional segments rather than treating the image as a single unit, the system achieves better recognition accuracy for objects with various postures while maintaining processing efficiency.
2Measurement precision
If multiple images are taken from different directions to improve recognition accuracy, then device complexity increases
Solution Approach 1:
Instead of using multiple cameras or imaging devices to capture images from different directions, the patent segments a single taken image into multiple partial images. This approach achieves multi-directional analysis capability while avoiding the complexity of multiple imaging devices, maintaining a simple single-camera configuration.
Solution Approach 2:
The patent creates multiple virtual viewing angles by dividing the single taken image into partial images and performing recognition on each segment. This virtual copying of the image into multiple analytical perspectives achieves the effect of multi-directional imaging without requiring physical multiple cameras or complex device arrangements.
3Measurement precision
If object posture is adjusted to optimize recognition, then recognition accuracy is improved, but ease of operation deteriorates
Solution Approach 1:
The patent eliminates the need for manual object posture adjustment by automatically dividing the taken image into partial images and performing recognition on each segment. This segmentation approach allows the system to automatically handle objects in various postures without requiring the operator to adjust the object's orientation, significantly improving ease of operation.
Solution Approach 2:
The system performs self-adjustment by automatically analyzing multiple partial images from a single taken image, effectively compensating for suboptimal object postures without human intervention. The automatic multi-directional analysis on segmented images allows the system to adapt to various object orientations independently, removing the burden of manual posture adjustment from the operator.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system enhances object recognition accuracy by leveraging images from different directions and reliability thresholds, ensuring high-precision identification of items on a register terminal with a simple configuration.
Implementation Method 1
a reflecting unit that is disposed within an imaging range of the camera and has a reflecting surface inclined with respect to an optical axis of the camera
Data Source
AI summary
An information processing system includes an imaging unit that generates an image signal by imaging and an information processing device. The information processing device performs at least any one of plural kinds of image processing on a taken image corresponding to the image signal. The information processing device specifies an object corresponding to a partial image included in the taken image on the basis of a state of the object corresponding to the partial image included in the taken image or a degree of reliability given to a processing result of the performed image processing.