Inference Model Weighting for Tracking Object Collation Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for collating objects, such as people, lack accuracy in determining matching analysis targets based solely on feature amounts, leading to suboptimal collation results.
Innovation Solution
A learning apparatus and method that generates ground truth weights for tracking object data using ground truth tracking object pair information and trains an inference model to output tracking object data weights, which are then used to calculate a collation score by associating similarity between tracking object data pairs, thereby improving collation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If collation is performed based solely on feature amounts, then the process is simple, but collation accuracy deteriorates
Solution Approach 1:
The collation process is segmented into multiple stages: feature amount extraction, weight generation by the inference model, and weighted similarity calculation. This segmentation allows each component to be optimized independently, improving overall accuracy without overwhelming complexity
Solution Approach 2:
The patent introduces weight parameters that dynamically change based on feature importance. By transforming the collation from a simple feature comparison to a weighted parameter-based evaluation, the system achieves higher accuracy while maintaining manageable complexity through the inference model
2Measurement precision
If ground truth weights are generated and inference model training is performed, then collation accuracy is improved, but device complexity increases
Solution Approach 1:
The inference model is trained in advance using ground truth weights before actual collation operations. This preliminary action separates the complex training phase from the operational phase, allowing the collation system to use pre-computed weights and reducing real-time computational complexity
Solution Approach 2:
The inference model acts as an intermediary component that bridges raw feature amounts and collation results. It transforms complex feature analysis into simplified weight outputs, reducing the complexity burden on the main collation system while maintaining high accuracy
Data Source
AI summary
A ground truth weight generation unit generates a ground truth weight for each piece of tracking object data of tracking object information regarding a tracking object by using ground truth tracking object pair information that is a set of tracking object information of the same tracking object or a set of tracking object information of separate tracking objects. An inference model training unit trains, by machine learning, an inference model that outputs a tracking object data weight corresponding to tracking object data included in the tracking object information by using data regarding the tracking object information as input data and using a ground truth weight generated for the tracking object information as ground truth data.