Image Inference Relearning Using Heat Maps to Correct Model Basis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing learned models often make inferences based on incorrect inference bases, necessitating a correction mechanism.
Innovation Solution
An information processing device that acquires an inference object image and a learned model, generates heat maps, extracts features, creates modification images, and relearns the model using relearning data to correct the inference basis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a learned model is used for inference, then inference speed and efficiency are improved, but the model may make inferences based on incorrect inference bases
Solution Approach 1:
The system generates heat maps to visualize the inference basis, compares it with correctly inferred heat maps, and feeds back correction information to the learned model through relearning processing. This feedback loop enables the model to correct its inference basis while maintaining fast inference capabilities.
Solution Approach 2:
Heat maps serve as an intermediary that bridges the learned model and the inference basis verification process. The heat maps visualize which regions the model uses for inference, allowing external verification and correction without slowing down the core inference process.
2Reliability
If heat maps are generated to verify inference basis, then inference basis correctness is improved, but processing time and computational load increase
Solution Approach 1:
The system generates heat maps selectively - primarily for verification purposes rather than for every inference. The heat map generation is performed as a partial action to verify critical inferences, balancing verification thoroughness with processing time constraints.
3Reliability
If relearning processing is performed to correct inference basis, then model accuracy is improved, but device complexity and processing steps increase
Solution Approach 1:
The learned model performs self-correction through the relearning processing unit. The system automatically generates correction data from heat map comparisons and feeds it back to the model, enabling self-service improvement without requiring complex external intervention systems.
Data Source
AI summary
An information processing device includes an acquisition unit that acquires an inference object image and a learned model, an inference unit that makes an inference by using the inference object image and the learned model, a generation unit that generates a heat map by using an inference result, an extraction unit that extracts a plurality of features based on the heat map, and a relearning processing unit. The generation unit generates a plurality of modification images by using the inference object image. The inference unit makes the inference by using the modification images and the learned model. The generation unit generates heat maps by using a plurality of inference results. When an inference basis is erroneous, the relearning processing unit generates relearning data, in which the feature indicated by the inference basis has been modified, by using learning data and relearns the learned model.