Representative Sample Selection for Accurate Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The selection of training samples for artificial intelligence models significantly impacts their training effectiveness and subsequent prediction accuracy, leading to low accuracy in object recognition tasks.
Innovation Solution
A method involving obtaining candidate object sets, performing feature extraction, clustering, and selecting representative objects based on recognition probabilities to improve model training and accuracy, using techniques like K-means and RALM for object category recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional training sample selection methods are used, then the model training process is simple, but the prediction accuracy is low
Solution Approach 1:
The patent segments the training sample selection process into multiple stages: initial sample selection, model training, accuracy evaluation, and iterative refinement. By dividing the complex selection process into manageable segments with specific criteria for each stage, the system achieves higher prediction accuracy while maintaining controllable process complexity
Solution Approach 2:
The patent implements a feedback mechanism where model prediction accuracy is continuously evaluated and used to guide subsequent sample selection. The system monitors performance metrics and uses this feedback to iteratively refine the training sample set, creating a closed-loop process that progressively improves prediction accuracy
2Adaptability or versatility
If more training samples are used, then the model coverage improves, but the resource consumption increases
Solution Approach 1:
The patent applies partial action by selecting only the most representative and informative samples for training, rather than using all available data. The system identifies and prioritizes samples that provide maximum learning value, achieving adequate model coverage with a subset of the total data, thereby reducing computational resource consumption while maintaining training effectiveness
Solution Approach 2:
The patent applies local quality by assigning different weights and priorities to different samples based on their representativeness and informational value. Rather than treating all samples uniformly, the system identifies high-value samples in specific regions of the data space and focuses training resources on these locally optimal samples, improving efficiency while maintaining comprehensive coverage
Data Source
AI summary
An object recognition method includes: obtaining a candidate object set having a plurality of candidate objects; obtaining a plurality of pieces of candidate object information of the candidate objects on a plurality of dimensions, performing feature extraction on the candidate object information to obtain candidate object features, fusing the candidate object features to obtain object extraction features corresponding to the candidate objects, and performing object category probability recognition based on the object extraction features to obtain a recognition probability that the candidate objects belong to a target object category; clustering the object extraction features to obtain sub extraction feature sets corresponding to clustering categories, and forming one sub object set from candidate objects corresponding to the object extraction features; and obtaining representative objects by selecting from the sub object sets respectively based on recognition probabilities corresponding to the candidate objects in the sub object sets.