Overlap-Based Multi-Model Image Recognition for Varying Object Sizes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image recognition technologies struggle with accurate detection of recognition target objects with varying image sizes and often produce erroneous results, particularly when using a single image recognition model for objects at different distances.
Innovation Solution
An image recognition apparatus and method that employs multiple image recognition models with different detection algorithms, performing multiple object detection processes to generate inference results based on the overlap and reliability of detected regions, and integrates these results to improve accuracy and suggest error likelihood.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single image recognition model is used for object detection, then the device complexity is reduced, but the measurement precision deteriorates for objects with varying image sizes
Solution Approach 1:
The patent combines multiple image recognition models (first image recognition model and second image recognition model) into a unified detection system. Each model processes the input image independently to generate object detection regions, and then these regions are merged through overlap evaluation to produce the final detection result. This merging approach allows the system to leverage the strengths of different models while maintaining manageable complexity through modular architecture.
Solution Approach 2:
The patent creates a universal detection framework where multiple image recognition models can be applied to detect objects with varying image sizes. The system evaluates overlap between detection regions from different models and uses this information to determine reliability, making the system universally applicable to objects at different distances and sizes without requiring model-specific optimization for each scenario.
2Measurement precision
If multiple image recognition models are used to detect objects with varying image sizes, then the measurement precision is improved, but the device complexity increases
Solution Approach 1:
The patent merges the output of multiple object detection processes by evaluating the overlap between detection regions. Instead of processing objects completely separately, the system combines results from the first and second image recognition models by comparing their detection regions, thereby reducing the effective complexity while maintaining high precision through multi-model validation.
Solution Approach 2:
The patent introduces a feedback mechanism where the overlap evaluation results are used to determine the reliability of detection. The system uses the overlap information to generate inference result data that indicates the likelihood of correct detection, providing feedback that can guide further processing or filtering of detection results to maintain precision while managing computational complexity.
Data Source
AI summary
In an image recognition apparatus, a processor performs, based on an input image and using an image recognition model, a plurality of object detection processes to detect as an object detection region a region in the input image where a recognition target object is judged to be present. In the plurality of object detection processes, a plurality of mutually different image recognition models are used. The processor generates inference result data according to the degree of overlap among a plurality of object detection regions detected in the plurality of object detection processes.