Digital watermarking of 3D disparity map based images for transaction tracking
Bibliographic record
Abstract
In this thesis, the basics of disparity map and watermarking are reviewed extensively. In order to embed binary information into images, a 3D image watermarking system was proposed. This embedded information was to survive the 3D Image rendering process of Disparity maps, to help identify malicious user who would distribute the watermarked image through an unauthorized system. The proposed system adopted the concept of hidden pixel and introduced an algorithm that identifies all known hidden pixels within the image. This information is combined with the Disparity map to generate a hidden pixel disparity map (HPDM); using the information in the HPDM a decision matrix is generated. This decision matrix is used to guide the watermark embedding process to ensure that information embedded in the Left Image can survive the 3D rendering process. Using the decision matrix, the watermark detector is capable of extracting the image from either the left or right image with no effect on the overall bit rate. This achievement is due to two original additions to the detection process: (1) Reverse rendering and (2) Cyclical Redundancy check. The proposed reverse rendering process expands the decision matrix into a reduced disparity map. This reduced disparity map is used to reverse the right image into a reduced left image. The identification of the image (left or right) is achieved through the use of a CRC check, which is also capable of detecting any errors in the extracted message, thus reducing the number of misidentification. The proposed system was implemented and tested using MATLAB. The bit efficiency of the proposed system varied between 38% and 88%. This variance is caused by the complexity of the depth scene as well as the cost function used in the depth estimation process. The watermark embedding system proposed had a PSNR of 45 dB (when no mark was embedded); this value is primarily attributed to some of the quantization that occurs during the DCT transform. However, a PSNR of 33dB is attained when the watermark was added at full strength.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".