Neural Network Training Through Object Grouping for Material Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for automatic and continuous detection and recognition of objects in object processing, such as recycling, lack efficiency, speed, and quality in material classification.
Innovation Solution
A method and system that utilize neural networks for object processing, including a camera, display, and processor to capture, group, and evaluate object images, employing unsupervised and semi-supervised learning tasks, with data adaptation routines to improve neural network training.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional object classification processes are used, then basic detection and classification can be performed, but precision in material detection and classification is insufficient
Solution Approach 1:
The system implements feedback mechanisms where classification results are continuously evaluated and used to refine the neural network models. The automated learning routine processes object sequences and feeds back improved classification parameters to enhance both detection precision and classification reliability iteratively
Solution Approach 2:
The system performs preliminary grouping of object images based on visual likeness before final classification. This preliminary action organizes data into structured sequences that improve the accuracy of subsequent detection and classification operations
2Measurement precision
If more sophisticated neural network models are deployed, then detection and classification quality improves, but system complexity increases
Solution Approach 1:
The system segments the classification process into distinct modular routines: object detection, image grouping, sequence evaluation, and automated learning. Each module performs a specific function with dedicated neural networks, improving classification quality while managing complexity through functional decomposition
Solution Approach 2:
The neural network models are designed to perform multiple functions: feature extraction, grouping, evaluation, and classification. This multi-functionality reduces the need for separate specialized systems, maintaining high classification quality without proportionally increasing overall system complexity
3Measurement precision
If continuous monitoring and evaluation are performed, then classification accuracy improves, but processing time increases
Solution Approach 1:
The system performs evaluation and learning tasks at periodic intervals using captured object sequences rather than continuously processing every single object in real-time. This periodic action maintains classification accuracy while reducing overall processing time and computational load
Solution Approach 2:
Object images are grouped and sequences are prepared in advance before final evaluation and classification. This preliminary organization of data structures accelerates the subsequent processing steps, reducing the time required for accurate classification
Data Source
AI summary
A method and a system for improving neural network training in object processing. The method comprises: receiving, from an object classification process using scene images of objects in an object processing facility, a set of segmented and classified objects captured during a pre-determined period; grouping the object images of the segmented and classified objects by a grouping routine based on their visual likeness to generate grouped object images, the grouping routine comprising at least one neural network; evaluating the objects of the object images based on comparison scores by a comparison routine, and generating a plurality of object sequences; and executing, by an automated learning routine, unsupervised and semi-supervised learning tasks by using the plurality of object sequences.