Interdependent Object Material Classifier Feedback Loop
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current object and material classification technologies, despite improvements, still result in a significant number of misclassifications, with a best case recognition rate of about 64%, indicating a need for enhanced recognition probabilities in both object and material classification.
Innovation Solution
The implementation of an interdependent object/material classifier with feedback mechanisms, utilizing a two-step classification process where the first phase employs known classifiers to obtain initial probabilities, and the second phase uses Bayesian algorithms and interdependent weighting functions to refine these probabilities, incorporating feedback between object and material classifications to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional object and material classifiers are used independently, then the classification process is simple and fast, but the recognition accuracy is limited to about 64%
Solution Approach 1:
The patent merges object classification and material classification into a unified interdependent system where both classification processes share computational resources and feedback loops. The object classifier and material classifier operate simultaneously with mutual influence, sharing the same feature extraction and probability calculation framework, thereby improving accuracy without proportionally increasing complexity.
Solution Approach 2:
The patent implements feedback mechanisms where the output probabilities from object classification are fed back into material classification, and vice versa. The interdependent weighting function uses feedback from both classification streams to iteratively refine the final probabilities, allowing each classification to benefit from the other's results and achieve higher recognition accuracy.
2Reliability
If independent object and material classification is used, then the system is easier to implement, but about 3 out of 10 objects are misclassified or unclassifiable
Solution Approach 1:
The feedback mechanism allows the system to iteratively refine classification results by using the output of one classification process as input for the other. This feedback loop continuously adjusts the probability assignments, enabling the system to correct misclassifications and improve reliability, especially for ambiguous cases where independent classification would fail.
Solution Approach 2:
The patent changes the parameters used in classification by introducing interdependent weighting functions that dynamically adjust the importance of different features and probabilities based on feedback from both object and material classification processes. This parameter adjustment allows the system to adapt to different object types and material characteristics, improving overall reliability.
3Measurement precision
If a two-phase classification process with feedback is implemented, then recognition probability is improved, but the classification time increases
Solution Approach 1:
The patent ensures continuity of useful action by running object and material classification processes simultaneously rather than sequentially, with both phases operating in parallel and exchanging feedback continuously. This concurrent execution maintains productive activity throughout the classification process, minimizing idle time while achieving high recognition probability through iterative refinement.
Data Source
AI summary
The present disclosure relates to classification of a material type of an object. A first phase applies an object classifier and a material classifier to obtain first object and material probabilities of the object. A second phase applies an interdependent object/material classifier to the first object and material probabilities to obtain further object and material class probabilities. The interdependent object/material classifier performs multiple iterations of calculating the further object and material class probabilities, and utilizes feedback in which an immediately preceding calculated prior further object class probability is included in a next iteration of calculating a further material class probability, and an immediately preceding calculated prior further material class probability is included in a next iteration of calculating a further object class probability.