Recognition Support Apparatus for Sorting System OCR Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing OCR processing systems in post offices and distribution centers face difficulties in obtaining correct solutions for image recognition tasks, leading to challenges in automatically adjusting parameters, procedures, and rules, resulting in incorrect recognition of delivery objects.
Innovation Solution
A sorting system that includes a delivery processing apparatus, a video coding terminal, and a recognition support apparatus with an information management portion, a correct solution derivation portion, and a machine learning portion, which acquires images and keying information to derive correct solutions and adjust OCR processing stages using machine learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If OCR processing is performed manually or with basic automation, then the system is simpler to operate, but the accuracy of delivery object recognition is insufficient
Solution Approach 1:
The OCR processing is divided into multiple distinct stages: image acquisition, preprocessing, character recognition, and post-processing. Each stage can be independently optimized and adjusted, allowing the system to achieve high recognition accuracy while maintaining manageable complexity through modular design.
Solution Approach 2:
The system incorporates dynamic parameter adjustment capabilities where processing parameters, procedures, and rules can be automatically tuned based on feedback from recognition results. This dynamic adaptation allows the system to optimize performance for different delivery objects without requiring complete system redesign.
2Measurement precision
If multiple processing stages are implemented in OCR, then recognition accuracy can be improved, but it becomes difficult to automatically adjust parameters, procedures, and rules
Solution Approach 1:
The system implements feedback mechanisms where recognition results are evaluated and used to automatically adjust parameters, procedures, and rules for subsequent processing. This closed-loop control enables automatic optimization of multiple processing stages without manual intervention, resolving the contradiction between multi-stage processing and ease of parameter adjustment.
Solution Approach 2:
The OCR system performs self-adjustment of its processing parameters and procedures based on its own performance metrics and feedback from recognition results. This self-service capability eliminates the need for manual tuning of multiple processing stages, making the system easy to operate while maintaining high accuracy.
3Measurement precision
If feedback mechanisms are added to adjust OCR parameters, then recognition accuracy improves, but the system complexity increases
Solution Approach 1:
The system pre-establishes the feedback mechanisms and adjustment rules during the design phase, so that during operation, the feedback loops automatically execute without adding operational complexity. The parameter adjustment procedures are prepared in advance and triggered automatically based on recognition performance, maintaining simplicity while improving accuracy.
Data Source
AI summary
According to one embodiment, a sorting system has a delivery processing apparatus, a video coding terminal, and a recognition support apparatus. The recognition support apparatus has an information management portion, a correct solution derivation portion, and a machine learning portion. The information management portion acquires an image obtained by imaging a delivery object and keying information. The correct solution derivation portion derives respective correct solutions for a plurality of processings for the image obtained by imaging the delivery object, based on the keying information acquired by the information management portion. The machine learning portion performs machine learning using the respective correct solutions for the plurality of processings derived by the correct solution derivation portion, to adjust the plurality of processings.