Embed-Image Generation via Block Segmentation and Positioning Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data embedding methods for print materials face issues where information blocks cannot be identified correctly, leading to incomplete data reading and reduced detection performance, especially in images with limited readable areas.
Innovation Solution
An image processing device and method that generates an embed-image by dividing information into blocks, creating information and positioning patterns based on block information, allowing for accurate identification and extraction of embed-information by using positioning patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If information is divided into blocks and block numbers are assigned for embedding, then data embedding capability is improved, but block identification reliability deteriorates when block numbers cannot be correctly read
Solution Approach 1:
The patent applies preliminary action by pre-arranging reference patterns at specific positions within each information block before embedding. These reference patterns serve as predetermined markers that enable reliable block identification during reading, even when block numbers themselves cannot be correctly read. The reference patterns are embedded in advance at known positions, allowing the reading device to locate and identify blocks without relying solely on block number accuracy.
2Ease of operation
If block number estimation is performed based on degree of agreement, then identification capability is improved, but measurement precision deteriorates when very similar data exists among information blocks
Solution Approach 1:
The patent extracts the identification function from the block number data itself and separates it into dedicated reference patterns. By taking out the identification purpose from the information-carrying block numbers, the system uses specialized reference patterns that are optimized solely for identification purposes. This separation allows accurate block identification even when information blocks contain very similar data, as the reference patterns provide unambiguous positional markers independent of the block content.
3Quantity of substance
If information blocks are arranged in document images, then data embedding capacity is improved, but detection performance deteriorates in images with limited readable areas
Solution Approach 1:
The patent applies segmentation by dividing the embed-information into multiple information blocks, each containing reference patterns at specific positions. This segmentation allows the system to process and identify blocks independently, so that even if only a limited area of the document is readable, the reference patterns in the visible portion can still be detected and used to identify those blocks. The segmentation into manageable blocks with embedded reference markers enables effective detection in limited readable areas while maintaining overall data embedding capacity.
Data Source
AI summary
It is possible to provide an image processing device, an image processing method, and its program which generate a positioning pattern of an information pattern in accordance with block information. This eliminates the problem that a block number cannot be identified although an information block position can be identified, i.e., the problem that information for identifying an information block cannot be sufficiently extracted. If an information pattern is identified, the extracted information can be used effectively. Moreover, it is possible to enhance the information embedding efficiency.