Image Inference ROI Mapping for Related Result Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques fail to grasp the relatedness between multiple inference results derived from identical input information.
Innovation Solution
An information processing apparatus and method that utilizes multiple inference units to generate different inference results from the same input data, determines the relatedness of focused regions of interest within the input data, and presents relatedness as reference information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If multiple inferences are made from identical input information, then the ability to present reference information is improved, but the capability to grasp relatedness between inference results is insufficient
Solution Approach 1:
The patent introduces regions of interest (ROIs) as intermediary elements to bridge multiple inference results. Instead of directly comparing inference outcomes, the system extracts ROIs from input information that correspond to each inference result, then determines relatedness based on the spatial and semantic relationships between these ROIs. This intermediary approach enables effective relatedness detection while preserving information for reference presentation.
Solution Approach 2:
The patent transforms the comparison problem from the inference result space to the region of interest space in the input information. By mapping inference results back to their corresponding regions in the original input data, the system creates a new dimension for comparison based on spatial inclusion relations and overlapping areas, making relatedness detection feasible and meaningful.
2Measurement precision
If regions of interest are extracted to determine relatedness, then the precision of relatedness determination is improved, but the complexity of the processing system increases
Solution Approach 1:
The patent segments the input information into multiple regions of interest, each corresponding to a specific inference result. This segmentation allows the system to focus computational resources on comparing specific regions rather than processing entire datasets, thereby improving determination precision while managing system complexity through localized processing.
Solution Approach 2:
The patent changes the parameters used for comparison from raw inference result data to derived region of interest parameters such as spatial coordinates, area overlaps, and inclusion relationships. This parameter transformation simplifies the complexity calculation by using geometric relationships that are more straightforward to compute and compare than raw inference data.
Data Source
AI summary
An information processing apparatus for making inferences using image data, comprising: a first inference unit configured to perform a first inference using the image data to obtain a first inference result; at least one second inference unit configured to make a second inference different from the first inference, using the image data, to obtain a second inference result; an information-of-interest acquisition unit configured to obtain a first region of interest which is region information focused in the image data, in the obtainment of the first inference result, and to obtain a second region of interest which is region information focused upon in the image data, in the obtainment of the second inference result; and a determination unit configured to determine relatedness of the first region and the second region of interest from an inclusion relation of the first region and the second region of interest in the image data.