OCR Pre-Optimization Using Inspector and Filter Sub-Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Optical Character Recognition (OCR) technologies face challenges in maximizing accuracy due to limitations in pre-capture image alteration and output text correction, especially when dealing with legacy datasets or restrictive capture environments.
Innovation Solution
A computer-implemented system that employs supervised-learning based inspector sub-models to predict transformation parameters for input data, followed by filter sub-models to derive convolution kernels, thereby optimizing the input data before further processing by an OCR engine.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If post-capture image transformations are applied to improve OCR accuracy, then text extraction accuracy is improved, but system complexity increases due to multiple processing stages
Solution Approach 1:
The system performs preliminary actions by applying post-capture image transformations before the OCR process. The inspector sub-model predicts transformation parameters and applies corrections to the input image, preparing it in advance for optimal OCR processing. This preliminary correction of distortions, rotations, and other image defects improves text extraction accuracy without requiring changes to the OCR engine itself.
Solution Approach 2:
The system introduces an intermediary component between the input image and the OCR engine. The inspector sub-model and filter sub-model act as intermediaries that process the image and generate optimized input for the OCR engine. This intermediary layer transforms the raw image into a pre-optimized format that maximizes OCR performance without modifying the underlying OCR system.
2Reliability
If multiple sub-models and transformations are used to optimize input data, then analysis performance is improved, but computational time and resources increase
Solution Approach 1:
The system performs preliminary optimization of input data before the main analysis process. By pre-processing images with the inspector and filter sub-models, the system prepares optimized input that reduces the computational burden during the actual OCR analysis. This preliminary action improves overall analysis performance while the computational overhead is concentrated in the pre-processing stage.
Solution Approach 2:
The inspector sub-model and filter sub-model automatically analyze and optimize input images without requiring manual intervention. The system self-adjusts transformation parameters and filter settings based on the specific characteristics of each input image, eliminating the need for manual tuning and reducing overall computational time through automated optimization.
Data Source
AI summary
Systems and methods for computer-implemented pre-optimization of input data before further processing thereof by a computer-implemented analyzation process, such as optical character recognition (OCR). A cooperative model is employed that combines one or more supervised-learning based inspector sub-models, and one or more filter sub-models that operating in series with the inspector sub-model(s). The inspectors first receive the input data and calculate one predicted transformation parameters then used to perform transformations on the input data. The inspector-transformed data is then passed to the filters, which derive respective convolution kernels and apply same to the inspector-transformed data before passing same to the OCR or other analyzation process. The inspectors may be pretrained with different training data. For OCR, the model is trained on minimization of normalized edit distance, the inspectors apply initial warping transformations to text images, followed by application of filtering transformations and input to the OCR engine.