Neural Network Image Data Extraction Architecture
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Extracting data from images, such as receipts, is challenging due to the complexity of accurately identifying and extracting relevant information from electronic images.
Innovation Solution
A neural network architecture is employed, involving a convolutional neural network to process images into a 3D data cube, followed by an encoder RNN to process slices, an attention layer for normalization, and a decoder RNN to generate probability arrays for character extraction, eliminating the need for optical character recognition (OCR).
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional OCR methods are used to extract data from images, then character recognition can be performed, but the process is complex and less accurate
Solution Approach 1:
The patent replaces traditional OCR mechanical character recognition processes with a neural network-based system. The convolutional neural network automatically learns and extracts features from images, substituting the manual, rule-based OCR approach with an adaptive, data-driven model that achieves higher accuracy while simplifying the overall extraction pipeline.
Solution Approach 2:
The patent transforms the image data through multiple parameter transformations in the neural network architecture. The convolutional layers transform spatial parameters, the recurrent layers transform temporal sequences of features, and the attention mechanism dynamically adjusts parameter weights to focus on relevant regions, enabling accurate data extraction without complex preprocessing.
2Productivity
If conventional data extraction methods are used, then processing can be performed, but efficiency and accuracy are limited
Solution Approach 1:
The patent performs preliminary feature extraction and image processing through convolutional layers before the main data extraction task. The neural network pre-processes the image by detecting edges, contours, and other relevant features, which accelerates the subsequent data extraction process while maintaining high accuracy through the integrated attention and recurrent processing mechanisms.
Data Source
AI summary
Embodiments of the present disclosure pertain to extracting data from images using neural networks. In one embodiment, an image is fit to a predetermined bounding window. The image is then processed with a convolutional neural network to produce a three dimensional data cube. Slices of the cube are processed by an encoder RNN, and the results concatenated. The concatenated results are processed by an attention layer with input from a downstream decoder RNN. The attention layer output is provided to the decoder RNN to generate a probability array where values in the probability array correspond to particular characters in a character set. The maximum value is selected, and translated into an output character. In one embodiment, an amount may be extracted from an image of a receipt.