DuDoCS-Net: dual-domain and self-attention based CycleGAN for low-dose SPECT myocardial perfusion image enhancement
Bibliographic record
Abstract
Objectives We aimed to develop a deep learning (DL) framework to predict standard-dose images from low-dose images for both reconstruction and projection domains in SPECT myocardial perfusion imaging (MPI). Methods We retrospectively enrolled 359 patients who underwent SPECT MPI, to estimate standard-dose data, using half-, quarter-, and one-eighth-dose data. A CycleGAN integrated with self-attention mechanism and skip connection was employed to restore standard-dose SPECT images at different dose levels for reconstruction and projection domains. Quantitative metrics (structural similarity index metrics, SSIM; peak signal-to-noise ratio, PSNR; root mean square error, RMSE) and clinical parameters were calculated to evaluate the image quality. We first evaluated the predicted image quality by conventional PSNR, SSIM and RMSE, based on which, clinical parameters including left ventricular volume (LV), end-systolic volume (ESV), end-systolic volume (EDV) and ejection fraction (EF) were then adopted to assess the clinical feasibility. A commercially available software (PMOD) was adopted to assess the diagnostic accuracy. Furthermore, a comparison experiment was conducted to explore whether projection-domain denoising is superior to the reconstruction domain. Results For non-gated cases, the lowest RMSE (1.56 ± 0.56) and highest PSNR (44.63 ± 1.69), SSIM (0.99 ± 0.01) were achieved in half-dose reconstructed images. For gated data, the lowest RMSE (2.18 ± 0.56) and highest PSNR (43.18 ± 1.69), SSIM (0.99 ± 0.01) were achieved. The proposed method achieved a performance that surpassed the advanced approaches with lowest RMSE (2.34 ± 0.48) and highest PSNR (42.07 ± 1.69), SSIM (0.99 ± 0.01), maintaining the fastest inference speed. The quantitative performance of the projection domain is superior to that of the reconstruction domain at different dose levels with lowest RMSE (half: 1.24 ± 0.42; quarter: 1.97 ± 0.46; one-eighth: 3.86 ± 0.62) and highest PSNR (half: 46.18 ± 1.67; quarter: 45.57 ± 1.52; one-eighth: 40.16 ± 1.63), SSIM (half:0.99 ± 0.01; quarter: 0.99 ± 0.01; one-eighth: 0.99 ± 0.01). The polar map analysis demonstrated the improved RMSEs (LAD: 2.02 ± 0.37; LCX: 2.15 ± 0.41; RCA: 2.02 ± 0.35). Conclusion The proposed method improves the quality of the estimated images in both reconstruction and projection domains. Projection-domain denoising surpasses the reconstruction domain for SPECT MPI.
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 | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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 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".