Business Card Array Detection via Character Distribution Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing devices struggle to accurately identify and separate business card images from a single read image containing multiple business cards, often leading to erroneous recognition of document types based solely on size, without considering the distribution of character strings and sizes.
Innovation Solution
An image processing device equipped with a document image extraction unit, a business card array image detection unit, and a business card image separation unit, which detects business card arrays by analyzing the distribution of character string objects and sizes within the document image, allowing for precise separation of business card images without relying on dictionaries or pre-registered phrases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If document type recognition is based solely on image size, then the recognition process is simple and fast, but the recognition accuracy deteriorates when multiple business cards are arranged without gaps
Solution Approach 1:
The patent segments the document image into multiple candidate regions by detecting character string distributions and boundaries. Instead of treating the entire image as one document, the system divides it into potential individual business card regions, then analyzes each region's characteristics separately to determine document types accurately even when cards are arranged without gaps.
Solution Approach 2:
The patent transitions from one-dimensional size-based recognition to multi-dimensional analysis by incorporating character string distribution patterns, character size variations, and spatial arrangement characteristics. This dimensional expansion enables accurate document type identification beyond simple image dimension comparisons.
2Measurement precision
If character string distribution analysis is performed to accurately detect business card arrays, then detection accuracy improves, but processing time and computational load increase
Solution Approach 1:
The patent performs preliminary character string detection and distribution analysis across the entire image before proceeding to detailed business card array detection. By pre-identifying character string locations and patterns, the system reduces the computational burden of subsequent analysis and accelerates the overall processing time while maintaining high detection accuracy.
3Reliability
If multiple analysis methods are combined to prevent erroneous recognition, then recognition reliability improves, but the processing complexity increases
Solution Approach 1:
The patent employs parameter changes by analyzing multiple characteristics of candidate regions including character string density, character size distribution, boundary patterns, and spatial relationships. By varying the analysis parameters and comparing results across different criteria, the system achieves reliable document type recognition while managing processing complexity through systematic parameter evaluation.
Data Source
AI summary
An image processing device includes a document image extraction unit, a business card array image detection unit, and a business card image separation unit. The document image extraction unit extracts one document image in a read image obtained by scanning one or a plurality of documents placed on a platen glass of an image reading device. When the document is a plurality of business cards arranged without gaps, the business card array image detection unit detects the document image as a business card array image in which a plurality of business card images are arranged without gaps, based on at least one of the distribution of character string objects and the distribution of character sizes in the document image. The business card image separation unit separates the detected business card array image into the plurality of business card images.