Multi-Predictor Influence Visualization for Concurrent Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques face difficulties in concurrently evaluating the behaviors of different predictors performing prediction processes, such as object detection, due to differences in detected subjects, estimated positions, and appearing ranges among prediction models, leading to inefficient comparison and analysis of detection results.
Innovation Solution
An information processing method that obtains prediction results from multiple predictors, determines the influence of input data on these results, and presents the influences side by side or by superposition, allowing for concurrent evaluation of predictor behaviors by combining and displaying prediction results based on overlap or positional relationships of detection frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple predictors perform prediction processes independently, then prediction accuracy can be improved through diverse detection capabilities, but it becomes difficult to concurrently evaluate and compare their behaviors
Solution Approach 1:
The patent combines multiple prediction results from different predictors into a unified presentation format. By integrating the detection results and their influences into a single visual interface showing side-by-side comparisons and superimposed displays, the system enables concurrent evaluation of multiple predictors without requiring separate analysis processes.
Solution Approach 2:
The patent introduces an information processing system as an intermediary that automatically processes, aligns, and presents the prediction results from multiple predictors. This mediator handles the complex task of comparing detection frames, positions, and influences, transforming multiple independent outputs into a unified evaluation view that facilitates easy comparison.
2Measurement precision
If detection results from different predictors are compared manually, then detailed analysis can be performed, but user trouble and time consumption increase significantly
Solution Approach 1:
The patent performs preliminary processing of prediction results by automatically aligning detection frames, standardizing result formats, and pre-calculating comparative metrics before presentation. This preliminary action prepares the data in advance, so when users view the results, the comparative analysis is already structured and ready, significantly reducing the time needed for manual comparison while maintaining deep analysis capability.
3Loss of information
If detection frames from different predictors are displayed separately, then each result can be clearly viewed, but efficient comparison and combination analysis becomes difficult
Solution Approach 1:
The patent merges detection frames from multiple predictors into a unified visual display that shows both individual results and combined analysis. The system superimposes detection frames, aligns them spatially, and presents them in a coordinated manner that preserves detailed detection information while enabling efficient visual comparison and combination analysis across predictors.
Data Source
AI summary
An information processing method performed by a computer includes: obtaining prediction results that are results of prediction performed by predictors on same input data; obtaining, for each of the prediction results, an influence that the input data had on the prediction result; determining, based on the prediction results, one or more combinations of the prediction results; and presenting, side by side or by superposition, using a presentation device, influences obtained for prediction results that are included in the prediction results and are in a same combination among the one or more combinations.