Compression Domain Video Watermarking via DCT Coefficient Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video watermarking methods in the compression domain are ineffective against recompression attacks and require significant computations, and existing methods in the spatial domain are inefficient due to the need to uncompress and recompress video, while also being weak against digital video interactive (DVI) capture attacks.
Innovation Solution
A method and apparatus for video watermarking in the compression domain that segments the video into spatial domains, calculates differences in DCT values to select an extraction domain, and inserts or extracts watermarks in a way that resists recompression and DVI capture attacks by identifying monotonic domains with minimal distortion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If watermarking is performed in the spatial domain by uncompressing and recompressing the video, then the watermark can be inserted and extracted, but the computational complexity increases significantly
Solution Approach 1:
The patent divides the compressed video into multiple slices along the vertical direction, where each slice contains a subset of DCT coefficients. This segmentation allows the watermark to be embedded only in the frequency domain coefficients of selected slices, avoiding the need to process the entire video frame by frame, thus reducing computational complexity while maintaining watermark embedding capability
Solution Approach 2:
The patent transforms the watermark embedding process from the spatial domain to the frequency domain by operating on DCT coefficients directly. This parameter change allows watermarking to be performed without uncompressing and recompressing the video, significantly reducing computational complexity while maintaining the ability to insert and extract watermarks
2Device complexity
If a watermark is directly inserted into compressed DCT coefficients, then computational requirements are reduced, but the method becomes ineffective against recompression attacks
Solution Approach 1:
The patent makes the watermark embedding process adaptive by dynamically selecting which DCT coefficient blocks to modify based on the video content and compression characteristics. The method adjusts the embedding strategy according to the specific compression parameters and video data, enabling the watermark to survive recompression operations while maintaining computational efficiency
Solution Approach 2:
The patent combines multiple watermark embedding techniques within the frequency domain, creating a composite watermarking approach that operates on DCT coefficients while incorporating temporal and spatial redundancy. This composite method provides resistance to recompression attacks by leveraging multiple domains of information within the compressed data structure
3Productivity
If conventional compression domain watermarking is used, then computation is quick, but the structure is very weak against video recompression
Solution Approach 1:
The patent segments the video into multiple slices and further divides each slice into multiple DCT coefficient blocks, allowing selective watermark embedding in specific blocks. This segmentation enables the method to maintain quick computation by processing only relevant blocks while building recompression resistance through distributed watermark placement across multiple slices and blocks
Solution Approach 2:
The patent performs preliminary analysis of the compressed video data to identify suitable DCT coefficient blocks for watermark embedding before actually embedding the watermark. This preliminary action allows the system to pre-determine the optimal locations that will provide both computational efficiency and strong resistance to future recompression operations
Data Source
AI summary
A video watermarking apparatus and method are provided, including an information provider for providing spatial domain segmentation information of an image with respect to a compressed video, a grouping unit for grouping the compressed video, based on the spatial domain segmentation information, and generating a grouped spatial domain, an extraction domain selector for selecting an extraction domain for extracting a watermark from the grouped spatial domain, and a watermark extractor for extracting a watermark bit inserted in the extraction domain.