Image Processing Device for Accurate OCR via Face Image Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing devices struggle to correctly determine the front and back sides of an image presenting medium, leading to incorrect optical character recognition (OCR) processing when the sides are incorrectly scanned, requiring manual re-scanning or correction.
Innovation Solution
An image processing device with a determination unit to detect the presence of face images in image data and a decision unit to decide the relationship between image data, allowing for swapping of image data output and adjusting processing accordingly, thereby ensuring accurate OCR on ID cards regardless of incorrect scanning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If image data is read from both sides of an image presenting medium without determining front and back sides, then reading speed is improved, but OCR processing accuracy deteriorates due to incorrect side identification
Solution Approach 1:
The determination unit performs preliminary detection of face images in read image data before OCR processing. By identifying the presence of face images (which typically appear on the front side of ID cards) in advance, the system determines which side is the front and which is the back, ensuring accurate OCR processing without requiring manual re-scanning.
Solution Approach 2:
The decision unit uses the determination result from the determination unit as feedback to adjust the processing flow. When a face image is detected, the system automatically swaps the image data to ensure correct front-back orientation before proceeding with OCR, creating a closed-loop control system that maintains accuracy while preserving high reading speed.
2Measurement precision
If manual re-scanning is performed to correct incorrect side identification, then OCR processing accuracy is improved, but processing time increases
Solution Approach 1:
The system performs automatic self-correction by detecting face images and automatically swapping image data when the front and back sides are incorrectly identified. This self-service mechanism eliminates the need for manual re-scanning or correction operations, maintaining high OCR accuracy while avoiding time loss associated with manual intervention.
3Measurement precision
If face image detection is performed for all image data, then side identification accuracy is improved, but device complexity increases
Solution Approach 1:
The determination unit extracts only the critical information needed for side identification - specifically, the presence or absence of face images - from the image data. By focusing on this single distinguishing feature rather than analyzing all image characteristics, the system achieves accurate side identification while keeping the determination unit relatively simple in structure.
Data Source
AI summary
An image processing device comprising a processor configured to: determine presence or absence of a face image for each of a plurality of pieces of image data read from at least one image presenting medium, and decide a relationship for the plurality of pieces of image data, based on a result of the determination. The relationship is a front and back of the image presenting medium or a page order, and the processor is further configured to: decides that image data including a face image is on a front side of the image presenting medium or on a page before image data not including a face image.