Adaptive Image Self-Recovery Based on Feature Extraction in the DCT Domain
Bibliographic record
Abstract
Image self-recovery aims at protecting digital images from partial damage due to accidental or malicious tampering. It is done by generating a reference code that contains the information of the image and embedding the code in the image itself. This code can later be extracted to restore the tampered regions of the image. The reference code must contain sufficient information to ensure a satisfactory reconstruction while being short enough to remain invisible when embedded in the image, which requires efficient extraction and adaptive encoding of the image information. To this end, we introduce a method for extracting local features in the DCT domain, in which the locations of the three DCT peaks, i.e., the DCT coefficients with the highest magnitudes, are examined to distinguish 13 texture profiles differing in the number of edges, edge orientations, and combinations of the two. Applying this method, we propose an adaptive image self-recovery algorithm. The DCT peaks are used to identify local texture patterns, and the bit allocation is made adaptive at hierarchical levels: 1) the texture blocks get more bit allocation than the smooth blocks; 2) the blocks having texture patterns appearing more frequently in the image are encoded with more precision; and 3) in each texture block, the highest DCT peak is assigned more bits than the remaining encoded coefficients. Hence, the encoding process is not only adaptive to the levels of variations across blocks but also to the local texture patterns. The proposed algorithm generates a reference code short enough to be embedded very comfortably in a single-least-significant-bit (LSB) plane, compared to 2 ~ 3 LSB planes often found in literature. Since the reference code contains all the critical image information in a compact form, the quality of the reconstructed images is as good as those produced by significantly longer reference codes.
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 | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 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".