Watermark Processing Device for High-Resolution Image Robustness
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing watermarking technologies face challenges in reducing processing time and calculation amount, especially when applying watermarks to high-resolution images from surveillance cameras, due to the need for multiple wavelet transformations, which are computationally intensive and time-consuming.
Innovation Solution
A method and device that create a watermarked image by wavelet-transforming a virtual image, inserting a watermark into its low-frequency sub-band, and inversely transforming it, allowing the watermarked image to be synthesized with the input image without transforming each frame of the input image, thereby reducing processing time and calculation load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a two-level wavelet transform is performed on the input image to insert a watermark, then the watermark insertion robustness is improved, but the calculation amount and processing time increase significantly
Solution Approach 1:
The patent divides the watermarking process into two separate stages: (1) generating a watermarked virtual image through wavelet transform and watermark insertion, and (2) synthesizing the watermarked virtual image with the actual input image. This segmentation allows the computationally intensive wavelet transform to be performed on a smaller virtual image rather than the full-resolution input image, significantly reducing processing time while maintaining watermark robustness.
Solution Approach 2:
The patent introduces a virtual image as an intermediary carrier for the watermark. Instead of directly watermarking the input image, the watermark is first inserted into the virtual image through wavelet transform, and then this watermarked virtual image serves as a mediator that is synthesized with the input image to produce the final watermarked output. This intermediary approach decouples the watermark insertion process from the input image processing, reducing computational burden.
2Reliability
If a two-level wavelet transform is performed on the input image to insert a watermark, then the watermark insertion robustness is improved, but the calculation amount increases
Solution Approach 1:
The patent segments the processing target by performing wavelet transform only on a virtual image with smaller dimensions rather than the full-resolution input image. This reduces the number of pixels subject to transformation, thereby decreasing the calculation amount while preserving the multi-level wavelet structure needed for watermark robustness.
Solution Approach 2:
The patent creates a virtual image as a copy or representation that carries the watermark information. This virtual image serves as a simplified model that undergoes the complex wavelet transformation and watermark insertion processes, while the actual input image remains unchanged until the final synthesis stage, reducing the overall computational complexity.
3Reliability
If watermark insertion is applied to high-resolution images from surveillance cameras, then the watermark visibility and robustness are improved, but the processing time and calculation amount become prohibitively large
Solution Approach 1:
The patent applies segmentation by separating the watermark insertion task from the full high-resolution image processing. The wavelet transform and watermark insertion are performed on a downsampled virtual image, while the high-resolution input image is only involved in the final synthesis stage. This segmentation enables watermarking of high-resolution surveillance images without prohibitive processing times.
Solution Approach 2:
The patent performs preliminary watermark insertion on a virtual image before combining it with the actual input image. By preparing the watermarked virtual image in advance through wavelet transform and watermark embedding, the system reduces the real-time processing burden when handling high-resolution surveillance footage, improving overall productivity.
Data Source
AI summary
Disclosed are a watermark processing device, a method thereof, and an image capturing apparatus capable of reducing a processing time and a calculation amount while inserting a watermark into an image. The watermark processing method includes creating a watermarked image by inserting a predetermined watermark into a virtual wavelet image; and creating a synthesized image by synthesizing an input image and the watermarked image.