Delivery Point Recognition Using Sequential ML Address Parsing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems for items like letters and parcels are limited by the time and resource intensity of optical character recognition (OCR) processes, especially when handling large volumes of items, which slows down the sorting and delivery processes.
Innovation Solution
A system and method using machine learning or deep learning models to recognize geographical area information, such as addresses, by building hierarchical databases and sequentially processing geographical area components without relying on OCR, significantly reducing processing time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical character recognition (OCR) processes are used to recognize delivery information, then recognition accuracy is maintained, but processing time increases significantly
Solution Approach 1:
The patent replaces the traditional OCR (optical character recognition) system with a machine learning-based image recognition system. Instead of using character-by-character recognition algorithms, the system uses trained neural networks to directly identify delivery information from images, substituting the mechanical/algorithmic OCR process with an AI-based approach that achieves both speed and accuracy
Solution Approach 2:
The patent changes the fundamental parameters of the recognition process by transitioning from text-based OCR to image-based machine learning. This involves changing the input data format, the processing algorithm, and the recognition methodology to achieve significantly faster processing times while maintaining accuracy
2Reliability
If traditional OCR methods are used for processing large volumes of items, then recognition reliability is maintained, but productivity decreases
Solution Approach 1:
The system replaces the traditional OCR mechanism with a machine learning-based recognition system that can process images much faster. The neural network models are trained to recognize delivery information patterns, enabling high-speed processing of large volumes of items while maintaining reliable recognition through the learned patterns and features
Solution Approach 2:
The patent implements preliminary action by pre-training machine learning models on large datasets of delivery information images before actual processing. This training phase allows the system to recognize patterns and make rapid predictions during operation, enabling high productivity without sacrificing reliability
3Productivity
If hierarchical database structure is implemented, then recognition efficiency is improved, but device complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the delivery information recognition task into hierarchical levels (e.g., country level, state level, city level, street level). Each level has its own specialized database and processing layer, allowing the system to handle different levels of detail separately and efficiently, improving overall recognition efficiency
Solution Approach 2:
The hierarchical database structure implements a nested organization where smaller, more specific databases are contained within larger, more general ones. For example, city-level databases are nested within state-level databases, which are nested within country-level databases, allowing efficient querying and processing at any level
Data Source
AI summary
This application relates to a system for automatically recognizing geographical area information provided on an item. The system may include an optical scanner configured to capture geographical area information provided on an item, the geographical area information comprising a plurality of geographical area components. The system may also include a controller in data communication with the optical scanner and configured to recognize the captured geographical area information by running a plurality of machine learning or deep learning models separately and sequentially on the plurality of geographical area components of the captured geographical area information.