Reconstruction of Compressive Sensed (CS) Images with Deep Equilibrium Model (DEQ) Based on Iterative Shrinkage-Thresholding Algorithm (ISTA)
Bibliographic record
Abstract
Compressive sensing (CS) is a signal processing technique that measures and compresses signal simultaneously and reconstructs it by finding solutions from underdetermined linear system. CS needs much smaller number of measurements than what is required by the Nyquist-Shannon sampling theorem. Traditional optimization algorithms require many iterations to reconstruct the signal and therefore it is time consuming. Accordingly, there is a need for quicker reconstruction methods. The objective of this thesis is to explore the application of neural network in CS inverse problem to get reconstruction quickly. This thesis has four main contributions. First, we use convolutional neural network (CNN) to improve efficiency of traditional CS reconstruction for sparse signal. The nonzero positions are directly found by CNN, then the rank deficiency model is converted into a full rank model and accurate reconstruction can be solved. Second, for the compressible signal such as image, we develop the semi-tensor product (STP) into a neural network. The measurement matrix is much smaller than that of traditional CS. The proposed STP-Net makes the sampling process convenient and provides good initial reconstruction without block artifacts. Third, inspired by iterative shrinkage-thresholding algorithm (ISTA), we build SPT-ISTA-Net. The model incorporates aggregated residual transformations (ResNeXt) to promote performance improvement and a squeeze-and-excitation (SE) block to enhance useful information. Fourth, to simplify the model structure, we built a deep equilibrium model dubbed as STP-DEQ-Net. The model uses one ISTA block as an implicit layer to implement an arbitrarily deep network. The model has competitive performance, and it has a trade-off between accuracy and computation. Multi-scale dilated convolutional layers are used to further improve performance.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".