Digital Watermark Detection via Spatial Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional digital watermark technologies face challenges in accurately identifying and extracting information when multiple sets of image data with embedded watermark information interfere with each other, leading to decreased detection accuracy and potential disfigurement of images.

Innovation Solution

A digital watermark detecting device that divides image data into areas with synchronized frequency and amplitude, where the phase is reversed based on the type of information, allowing for the identification and extraction of specific overlapping areas.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple sets of image data with embedded watermark information are projected simultaneously, then information embedding capability is improved, but detection accuracy deteriorates due to interference between watermark sets

Engineering Contradiction:
Improveinformation embedding capabilityVSAvoiddetection accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent divides the projection area into multiple distinct regions, with each region containing only one set of watermark information. This spatial segmentation prevents interference between different watermark sets by ensuring that each detector processes only a single watermark type, thereby maintaining high detection accuracy while enabling multiple watermarks to be projected simultaneously across different areas.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If conventional frame superimposition is used to define target areas, then area identification is improved, but image quality deteriorates due to disfigurement

Engineering Contradiction:
Improvearea identification accuracyVSAvoidimage disfigurement
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent extracts and removes the frame superimposition method that causes disfigurement, replacing it with a virtual frame concept defined through coordinate information. The target areas are identified by calculating coordinate ranges based on projector and detector positions without actually superimposing visual frames, thus eliminating image disfigurement while maintaining accurate area identification.

Inventive Principle:
Principle #2Taking out (Extraction)

3Manufacturing precision

If small area division is used for bit determination, then information extraction granularity is improved, but detection reliability deteriorates when watermark strength is weak

Engineering Contradiction:
Improveinformation extraction granularityVSAvoiddetection reliability
Core Design Contradiction:
Manufacturing precisionVSReliability

Solution Approach 1:

The patent combines multiple small area measurements into a unified detection approach by accumulating luminance variations across multiple frames for each small area. This merging of temporal data from multiple frames enhances the reliability of detection for weak watermarks, while the spatial granularity of small area division is preserved for precise information extraction.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP3171584B1Electronic watermark information detection apparatus and electronic watermark information detection method
Publication Date: 2021.08.11 FUJITSU LTD
  • EP3171584B1 patent drawingFigure 1
  • EP3171584B1 patent drawingFigure 2
  • EP3171584B1 patent drawingFigure 3~4

AI summary

A detecting device (200) includes a dividing unit (260b), an identifying unit (260c), and an extracting unit (260d). The dividing unit (260b) divides each set of image data, which is included in visual information (250a), into a plurality of areas. The visual information (250a) represents information of videos of information projected by lighting devices that project image data having a plurality of sets of digital watermark information, in each of which the frequency and the amplitude are constant and the phase is reversed depending on whether "bit 1" or "bit 0" is indicated, embedded therein in a synchronized manner. The identifying unit (260c) extracts the amplitude of the digital watermark information based on the small areas of each set of image data, and identifies such successive areas in which the amplitude of the digital watermark information is constant as the target areas for extraction. The extracting unit (260d) extracts information corresponding to digital watermarking information from the target areas for extraction.