Tamper Detection in Compressed Video via Luminance Watermarking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital authentication techniques using hash functions are ineffective in detecting tampering of video data after compression, as they require additional information that cannot be reprocessed and causes issues in data handling and distribution.
Innovation Solution
A data processing apparatus and method that encryptically processes digital video data by calculating representative values for pixel blocks, rewriting luminance data to ensure quantized integers are multiples of a preset integer, and applying orthogonal transform to produce a coded video signal, allowing for tamper detection even if information for finding tempering was added before compression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If hash function is used for digital authentication, then tampering detection capability is improved, but additional information must be added after compression which causes data handling issues and prevents reprocessing
Solution Approach 1:
The patent applies preliminary action by embedding tamper detection information (watermark) into the video data BEFORE compression processing. This is achieved by modifying pixel values in the original video data to encode detection information, which then survives the compression process. The watermark is prepared in advance rather than added after compression, resolving the contradiction by enabling both tamper detection and maintaining data reprocessing capability.
2Reliability
If information for finding tempering is added before compression, then data reprocessing is prevented, but this causes problems in data handling and distribution
Solution Approach 1:
The patent uses parameter changes by modifying pixel value parameters in a controlled manner to embed watermark information. The watermark is encoded by changing pixel values slightly (e.g., adding small offsets) while maintaining the visual quality and allowing subsequent compression and reprocessing. This parameter modification approach enables both tamper detection accuracy and data adaptability for reprocessing.
3Productivity
If irreversible compression is applied to video data, then transfer time is reduced, but tamper detection becomes impossible if authentication information is added before compression
Solution Approach 1:
The patent applies preliminary action by embedding the watermark BEFORE irreversible compression. The watermark is encoded into the pixel data in advance, and then the compressed video data is distributed efficiently. During verification, the same decompression process is applied and the watermark is extracted to detect tampering. This resolves the contradiction by enabling both fast transfer through compression and reliable tamper detection through pre-embedded watermarking.
Data Source
AI summary
Digital video data is encryptically processed and then subjected to an encoding procedure with orthogonal transform to produce a coded video signal. A representative value of luminance of pixels of the digital video data is obtained for each of pixel blocks having a predetermined number of pixels, each pixel block corresponding to a unit of data subjected to the encoding procedure, as a quantized integer for each pixel block by using a specific function with the stored luminance data as a parameter. It is determined whether the quantized integer of the representative value for each pixel block is a multiple of a preset integer N of two or larger, and, if not, the stored luminance data are rewritten with a specific requirement giving the least range of change in gradation so that the quantized integer becomes a multiple of the integer N, whereas the stored luminance data remain unchanged if the quantized integer is a multiple of the integer N, thus the digital video data being encryptically processed. The encryptically processed digital video data is encoded with orthogonal transform to produce a coded video signal. The coded video signal is decoded. A second representative value of luminance data of the decoded signal is obtained as a quantized integer for each pixel block by using the specific function with the luminance data as the parameter. The quantized integer of the second representative value is divided by the integer N. It is determined that the video data has been tampered with when remainders of division meet predetermined tamer criteria.