Bill Identification Using Area and Character Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Financial staff face challenges in automatically identifying payment amounts on invoices, which are crucial for efficient processing but often require manual handling due to the complexity and variability of invoice formats.
Innovation Solution
A method utilizing pre-trained area and character identification models to automatically identify the final payment area and amount on invoices, with optional segmentation and display of the results, and calculation based on expense summary areas using a preset rule base when the primary area cannot be identified.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual handling of invoices is used to identify payment amounts, then accuracy can be maintained through human judgment, but processing efficiency deteriorates due to the large volume of invoices requiring daily handling
Solution Approach 1:
The patent replaces the mechanical human judgment system with an automated image processing system comprising an area identification model and a character identification model. The area identification model locates the payment area on the invoice image, and the character identification model recognizes the payment amount characters within that area, thereby substituting manual handling with automated computational processing while maintaining both efficiency and accuracy
Solution Approach 2:
The patent introduces an intermediary area identification model that bridges the gap between the raw invoice image and the character recognition process. This intermediary model first identifies and delimits the payment area, which then serves as the focused input for the character identification model, enabling accurate extraction of payment amounts without requiring full invoice processing
2Productivity
If automated identification methods are implemented to improve processing efficiency, then productivity increases, but identification accuracy deteriorates due to the complexity and variability of invoice formats
Solution Approach 1:
The patent segments the invoice processing task into two distinct stages: first, the area identification model segments the invoice image to locate the payment area; second, the character identification model processes only the segmented payment area to recognize the amount. This segmentation approach handles format variability by focusing on the specific payment region rather than the entire invoice, thereby maintaining accuracy while enabling automated high-volume processing
Solution Approach 2:
The patent applies local quality by training the area identification model to specifically recognize the payment area's unique characteristics and the character identification model to recognize payment amount characters with specialized features. This localized specialization enables the system to handle diverse invoice formats by adapting to the specific visual properties of payment-related regions and characters
3Speed
If pre-trained models are used for automatic identification, then processing speed improves, but adaptability to different invoice formats deteriorates due to fixed model structures
Solution Approach 1:
The patent achieves universality by designing the area identification model and character identification model to handle multiple invoice formats through common visual patterns. The models are trained to recognize payment areas and amount characters across different invoice types (receipts, bills, invoices) by focusing on universal characteristics such as numerical patterns, currency symbols, and spatial relationships, enabling fast processing across diverse formats without requiring format-specific models
Data Source
AI summary
The disclosure provides a bill identification method, device, electronic device and a computer-readable storage medium. The method includes: obtaining an image of a bill to be identified; using a pre-trained area identification model to identify a final payment area of the bill in the image; using a pre-trained character identification model to identify the final payment amount in the final payment area. By applying the solution provided by the present disclosure, it is possible to realize automatic identification of the payment amount on the bill, and the efficiency of bill processing is improved.