Supplier Identity Prediction for Scanned Image Documents
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine learning models struggle to accurately predict supplier identities in image-type documents when the supplier information is not explicitly included in textual form, often due to noise introduced by scanning and the presence of logos and graphics that hinder text recognition.
Innovation Solution
A system generates non-textual and textual features from image-type documents to predict supplier identities, using a machine learning model trained on historical documents to classify scanned vs. non-scanned documents and extract key-value pairs, employing algorithms like neural networks and backpropagation to improve prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If text recognition applications are used to extract supplier information from scanned documents, then text extraction can be automated, but accuracy deteriorates due to noise from scanning and graphics/logos
Solution Approach 1:
The system segments the document processing into multiple stages: first extracting all text content, then identifying key-value pairs, and finally predicting missing supplier information using machine learning. This segmentation allows each stage to specialize and improve overall accuracy despite automated processing.
Solution Approach 2:
The machine learning model acts as an intermediary between the extracted text and the final supplier identification. It bridges the gap by inferring missing supplier information from contextual clues in the document when direct extraction fails, thereby maintaining automation while improving accuracy.
2Measurement precision
If manual scanning of invoices is performed to identify supplier information, then accuracy can be maintained, but productivity deteriorates due to time-consuming processes
Solution Approach 1:
The system performs preliminary text extraction and key-value pair identification using automated methods for all documents. Only when these preliminary actions fail to identify supplier information does it invoke the machine learning model, combining speed with accuracy through selective processing.
Solution Approach 2:
The machine learning model enables the system to self-correct automated extraction failures. When the initial automated process misses supplier information, the ML model independently analyzes contextual patterns and fills in the missing data without human intervention, maintaining both speed and accuracy.
3Extent of automation
If machine learning models are trained to extract content from image-type documents, then automation is improved, but reliability deteriorates when supplier fields are not explicitly identifiable
Solution Approach 1:
The system dynamically adjusts its processing approach based on document characteristics. It first attempts standard automated extraction, then selectively applies machine learning prediction only when supplier information is not explicitly identifiable, making the automation adaptive and reliable across varying document types.
Solution Approach 2:
The system incorporates feedback loops where the machine learning model's predictions are continuously refined based on their success in accurately identifying suppliers. This feedback mechanism improves the reliability of automated processing over time by learning from past performance on diverse document types.
Data Source
AI summary
Techniques for predicting a missing value in an image-type document are disclosed. A system predicts the identity of a supplier associated with an image-type document in which the supplier's identity may not be extracted by text recognition. When a system determines that the supplier identity cannot be identified using a text recognition application, the system generates a set of machine learning model input features from features extracted from the image-type document to predict the supplier's identity. One input feature is a data file bounds feature indicating whether the image-type document is a scanned document or a non-scanned document. The system predicts a value for the supplier's identity based on the data file bounds value and additional feature values, including color channel characteristics and spatial characteristics of regions-of-interest. The system generates a mapping of values to defined attributes based in part on the predicted value for the supplier's identity.